Warrior0302/CAP-Bench

1task_id,instruction
2task-035e06,"I’m writing a sci-fi script about AI ethics and want to systematically gather creative source material. First, go to IMDb and search for sci-fi films with theme tags including “artificial intelligence” or “dystopia,” with ratings above 7.5. Select 5 classic works and record the title, director, IMDb rating, and detail-page link for each.3 4Then, search these 5 films on Letterboxd. For each film, check the community rating and top user reviews (prioritize highly liked reviews; if the page cannot be reliably sorted by likes, select the 3 most relevant visible reviews and state your selection criteria), and extract keywords describing commonly discussed narrative techniques (e.g., “multi-thread narrative,” “time loop,” “first-person perspective”).5 6Next, on Goodreads, try to find the original novels corresponding to these films or books from the same IP/franchise. Prioritize books with ratings above 4.0 and more than 1,000 reviews, and aim to find 3 books. If fewer than 3 are available, keep the results found and clearly state the gap. Record the title, author, Goodreads rating, philosophical topics mentioned in popular reviews (e.g., “free will,” “the nature of consciousness”), and link.7 8Then, go to Semantic Scholar and search for “AI ethics” or “machine consciousness.” Sort by citation count and find 5 papers with more than 100 citations (if fewer than 5 exist, record the actual number found and explain). Record the title, authors, citation count, publication year, abstract, and paper link.9 10Finally, on YouTube, find in-depth analysis videos for any 2 of the 5 films (e.g., “[Film] review,” “[Film] analysis”), with each video longer than 15 minutes and having more than 100,000 views; find 1 video per film. If a specific film has insufficient results, you may switch to another film among the 5 and expand search keywords (e.g., “ending explained,” “deep dive”). Record the video title, channel name, view count, duration, and link.11 12Output format:13- IMDb films (title, director, rating, link)14- Letterboxd info (community rating, narrative-technique keywords)15- Goodreads books (title, author, rating, philosophical topics, link)16- Papers (title, authors, citation count, year, abstract, link)17- YouTube videos (title, channel, views, duration, link)18 19If any step encounters a login wall or regional restriction, record the limitation and continue with the remaining steps."
20task-070c65,"I’m preparing teaching materials for an Ancient Greek mythology art workshop. First, go to the Getty Museum website and find up to 5 Ancient Greek mythology-themed works that are currently on display (sculptures, ceramics, or reliefs are all acceptable). For each work, record its title, date, material, exhibition name, and the mythological figure or theme mentioned in its description (e.g., Athena, Heracles, the Trojan War, etc.). If there are fewer than 5 relevant works in Current exhibitions, supplement from the Getty collection database with additional Ancient Greek mythology-related works, and clearly mark which ones are from Current exhibitions and which are from the collection database.21 22Then go to Bilibili and YouTube, and use the names of these mythological figures as search keywords for educational/explanatory videos:23- On Bilibili: filter within the Knowledge category, find one video that is at least 8 minutes long with the highest view count.24- On YouTube: filter for videos at least 8 minutes long, sort by view count, and find the one with the highest views.25 26Finally, output for each artwork:27- Artwork title  28- Date  29- Material  30- Mythological theme  31- Getty artwork detail page link  32- Bilibili video title and link  33- Bilibili video duration and view count  34- YouTube video title and link  35- YouTube video duration and view count"
36task-0aae15,"I want to contribute code to the pandas Python library but don't know where to begin.37 38First, help me find open issues in the pandas project on GitHub that are labeled 'good first issue'. From those, identify the top 5 most active issues (those with recent updates and many comments), and for each, record its title, a brief summary of the topic, and the number of discussions.39 40Next, search for pandas-related questions on StackOverflow, sorted by votes, to find the most frequently asked and highly upvoted ones. List the titles and vote counts for the top 10.41 42Finally, compare these high-frequency StackOverflow questions with the themes of the GitHub issues. Identify issues that are suitable for beginners and simultaneously address significant user pain points. Recommend 2-3 of the most worthwhile issues for me to contribute to."
43task-0c0dd6,"I’m finalizing my itinerary for the next CES (Consumer Electronics Show). First, please check Wikipedia to confirm the exact dates, host city, and main venue name for the next CES. After obtaining the dates and city, search Eventbrite for “Tech Networking” or “After Party” events taking place in the same city during that period. Please identify 3 events that meet the following criteria: priced at $50 or less (or free), and held in the evening during CES dates (after 18:00). If fewer than 3 results are available during CES, you may extend the search to one day before and one day after CES, and clearly note this in the output. Once identified, use Google Maps to calculate the driving distance and estimated travel time from the CES main venue to each of the three event locations. Output the CES start and end dates, host city, and main venue name; and for each of the 3 events: name, exact time, price, Eventbrite link, and driving distance and time from the main venue."
44task-0e526a,"I want to reproduce image segmentation experiments. First, go to Semantic Scholar and search for image segmentation papers published since 2024 in this field. Identify up to 5 papers with relatively high citation counts, and keep only those with downloadable PDFs. Prioritize papers whose PDFs are directly accessible on a public webpage; if fewer than 5 meet this condition, proceed with however many are actually found.45 46Then, pick one of these papers and go to GitHub. Search by the paper title or the first author’s name to find a related code repository (official or community implementation). Prefer repositories with more than 500 stars. Verify that the README cites the paper and that the repository includes a `requirements.txt` file (if not, move to the next paper until one is found).47 48Next, go to Hugging Face and search for the paper title. Find up to 3 pretrained models with relatively high download counts, and check each Model Card to confirm the model actually corresponds to that paper; if fewer than 3 are available, record the actual number found.49 50Finally, go to YouTube and search by the paper title or first author. Find in-depth explainer videos longer than 10 minutes with more than 5,000 views. Expand the video description and confirm it mentions the paper title or author. If too few videos strictly meet all conditions, keep videos with duration ≥10 minutes and prioritize those with higher view counts and descriptions that clearly mention paper-related information.51 52Output:53- **Paper**: title, first author, publication year, Semantic Scholar citation count, whether PDF is available, paper link  54- **GitHub**: repository name, star count, whether `requirements.txt` is present, repository link  55- **Hugging Face**: model name, download count, Model Card summary, model link  56- **YouTube**: video title, duration, view count, video description summary, video link"
57task-109c02,"As a first-year graduate student, my advisor has tasked me with researching the 'LLM Watermarking' field, but I'm concerned about running into pitfalls. Please help me find a paper on Google Scholar published after 2023 with a relatively high citation count (at least 50 citations). Crucially, I need one that comes with runnable code. After finding a suitable paper, please navigate to its detail page and look for a GitHub link (typically located next to the PDF link or within the description). If one exists, click on the GitHub repository and briefly review the 'Open Issues' in its issue list. Specifically, look for unaddressed error reports such as 'Installation failed' or 'Environment error'. Please select the repository that appears to be the best maintained and has the fewest complaints. Then, provide me with its GitHub URL, the paper's title, and the current number of open issues."
58task-11ffd9,"As an avid NBA 2K player, I feel Klay Thompson's recent three-point rating in the game is too low, and I intend to post on a forum to discuss this.59 60First, please check IGN for recent news on NBA 2K25 player rating updates or find relevant top-player rating articles. Confirm Klay Thompson's current three-point rating and note if the articles mention any changes to his numerical value.61 62Next, go to Basketball-Reference to retrieve his real-world statistics from his most recent 20 regular season games. I specifically need his '3-point field goal percentage (3P%)' and '3-point attempts per game (3PA)'. It would be even better if you could find shot charts or zone shooting percentages for these 20 games. This information will help me ascertain whether his performance has genuinely declined or if the game's rating is unwarranted.63 64Finally, synthesize the data from both sources and inform me whether his in-game rating accurately reflects his recent real-world performance."
65task-1238b0,"Our family is having a dinner gathering next Saturday to Sunday. My sister has a severe nut allergy, and my cousin is allergic to shellfish. I need help finding some safe recipes.66 67First, please consult Mayo Clinic or WebMD to research cross-reactive foods and common hidden sources for both nut and shellfish allergies (e.g., nut ingredients in certain sauces or condiments). Compile and list these risky ingredients.68 69Next, navigate to AllRecipes and search for recipes explicitly labeled as ""nut-free"" and ""shellfish-free"". Identify 5 recipes that have a rating of 4 stars or higher and are suitable for a family dinner.70 71Subsequently, meticulously examine the ingredient list of each of these 5 selected recipes for any potential hidden risks. For example, certain sauces or seasonings might contain nut oils, tahini, or other allergenic components. If the safety of any specific ingredient within a recipe is uncertain, use Wikipedia or Healthline to investigate its detailed composition.72 73Finally, provide the following structured output: For each recipe, include its name, a list of its primary ingredients, its rating, the direct AllRecipes link, and your comprehensive allergy risk assessment. In your assessment, clearly indicate which ingredients have been verified as safe and which ones will require careful inspection of product labels during grocery shopping.74 75Additionally, furnish the complete lists of cross-reactive allergens and hidden sources for nut and shellfish allergies that you retrieved from the medical websites (Mayo Clinic or WebMD), along with the corresponding direct page links to these resources."
76task-13f79e,"I'm a jazz enthusiast currently in London, planning to go to New York to catch a live performance. Please first go to Eventbrite and help me find jazz performances in New York that still have tickets available for a weekend in the coming weeks, for example, from next Friday to next Sunday. Don't just pick the first results; scroll down or browse a few pages. Choose an event that looks interesting and still has tickets available. Note its name, specific performance time, and venue.77 78Based on this performance time, go to Skyscanner and search for round-trip flights from London (LHR) to New York. I want direct flights only; if none are available, choose the flight with the shortest travel time.79 80Finally, go to Expedia and look for hotels in New York for those performance days. Find one with a rating of 4 stars or higher that includes free breakfast.81 82Finally, compile and present the selected event, flight information, and hotel recommendation."
83task-1533cc,"I am planning to attend a 3-day conference in Miami, Florida, during the last week of January next year (specifically from the 28th to the 30th of January), and have already booked a stay at the Hilton Miami Downtown. However, I just saw news reports indicating there might be a hurricane passing through during those dates. I want to confirm the level of risk and whether I should reschedule or change location if the risk is significant.84 85Please first check AccuWeather for the weather conditions in Miami during the last week of January. If there are severe weather alerts (Watches or Warnings) or hurricane tracking indicates a storm might affect the area, help me assess the risk level. Then, check Booking.com for Hilton Miami Downtown's cancellation policy (free cancellation deadline and cancellation fees). If the policy allows and the risk is high, please search on Booking.com for alternative hotels in safer areas surrounding Miami (e.g., 50-100 miles north of Miami) for the last week of January, with the following requirements: a rating above 8.0, free cancellation support, and a price range of $150-$300 per night.86 87If the risk in Miami itself is low but flights might be affected, check Google Flights for flights to Miami departing on the 27th of January next year. Find 3 alternative flights (direct or one-stop, including different airlines and departure times), and record the flight numbers, departure/arrival times, fares, and rebooking policies.88 89Finally, if a hotel change is necessary, provide links to the official booking pages of the alternative hotels. If alternative flights are needed, provide links to the airline's official flight rebooking pages.90 91Output:921) AccuWeather's weather alert level (if any), alert type, affected period, and risk assessment;932) Hilton Miami Downtown's free cancellation deadline, cancellation fees, and Booking.com details page link;943) If alternative hotels are required, provide the names, addresses, ratings, per-night prices, free cancellation policies, Booking.com details page links, and official website booking links for 3 suitable hotels;954) If alternative flights are required, provide the flight numbers, airlines, departure/arrival times, flight durations, fares, Google Flights links, and airline official rebooking page links for 3 alternative flights;965) Contingency plan recommendations based on the above information (continue as planned / change hotels / reschedule)."
97task-16815d,"I am a film student in the directing department, currently researching the visual style of Film Noir. Please help me compile a shot-by-shot analysis research index.98 99First, go to the official Criterion website, filter for films categorized as ""Film Noir,"" and ensure they have a Blu-ray version. From these, select 3 classic representative works released between the 1940s and 1950s (preferably by renowned directors).100 101Then, go to IMDb to find the technical specifications for these 3 films. Specifically, I need their ""Aspect Ratio"" and ""Negative Format.""102 103Finally, go to Bilibili and search for in-depth shot-by-shot analysis videos of these 3 films. Find high-quality videos with over 10,000 views and titles that include ""拉片"" (shot-by-shot analysis) or ""解析"" (analysis/breakdown).104 105Output: Film Title, Director, Year, Criterion Link, IMDb Aspect Ratio, IMDb Negative Format, Bilibili Analysis Video Title, Views, and Link."
106task-17cf2c,"I'd like to plan a team-building event for my team of 20 people in San Francisco for an upcoming weekend.107 108First, on Eventbrite, search for an event in San Francisco, applying the 'This Weekend' filter to specify the date range for an upcoming weekend. Prioritize creative workshops or tours. Avoid events marked 'Sales end soon' or 'Going fast', and ensure the time is suitable.109 110Once an activity is selected, note down its specific address.111 112Then, go to OpenTable and search for restaurants within walking distance (approximately 15 minutes) of the chosen activity's location. I'm looking for a well-rated restaurant (4 stars or higher or 'High Rating') that can accommodate a party of 20 (Party Size: 20). Preferably, the cuisine should be Italian or American.113 114Finally, please provide me with the name, time, and location of the selected activity, along with the matching restaurant's name, rating, and specific available reservation times (e.g., 6:30 PM)."
115task-194f6d,"I found a data science industry summit in San Francisco on Eventbrite. I want to identify which companies associated with this event are currently hiring for data science-related roles. First, extract the company list from the event details page (possibly from sponsors, partners, organizers, agenda descriptions, or logo walls). If the event page does not explicitly provide an “exhibiting companies” field, extract all identifiable company/institution names on the page and label the section each one comes from.116 117Then, search LinkedIn and Indeed company by company to check whether they are hiring for **Data Scientist** or **Machine Learning Engineer** roles, prioritizing positions in **San Francisco** or **Remote**, and prioritizing jobs posted in the **last 30 days**. If results are very limited under these strict filters, keep the same company and role keywords but relax the posting window to the **last 60 days**, and clearly mark this in the results.118 119For each role found, record:120- Job title  121- Company  122- Location  123- Salary range (if available)  124- Key requirements (skills, years of experience, etc.)  125- Job link  126- Source platform127 128If you encounter LinkedIn login walls or access restrictions, record **“Restricted—unable to view details”**, continue with the remaining companies using whatever is visible, and use Indeed to supplement and cross-validate.129 130Finally, compile everything into a checklist indicating:131- Which companies have matching openings  132- Which companies currently have no matching openings  133 134This will help me decide whom to prioritize networking with at the event and prepare targeted questions and resumes."
135task-19636e,"I'm looking to purchase a highly replayable JRPG on Steam during the ongoing sales, with a budget of under 100 RMB. Please help me filter games on Steam that are tagged 'JRPG', priced under 100 RMB, and have a review rating of 'Very Positive' or higher. To ensure the gameplay isn't too short, please browse through several pages of the search results, select 3 promising titles, and then check their 'Main Story' completion times on HowLongToBeat. Finally, output a table listing the names of these 3 games, their current Steam price, review rating status, and Main Story length. Also, tell me which game offers the best 'price-per-hour' value."
136task-198aa0,"I'd like to get away for a short break, departing from New York (JFK) next Friday and returning on Sunday.137 138First, please use the ""Explore everywhere"" feature on Skyscanner, only looking for direct flights, to find the 3 cheapest specific destination cities in different countries (choosing the cheapest city from each country).139 140Once these 3 cities are decided, go to TripAdvisor to find a restaurant for each city. The requirements are: it must specialize in local cuisine (e.g., if going to Mexico, look for tacos), and have a rating of 4.5 stars or higher. Most importantly, please open and check the most recent reviews to ensure no one is complaining that it's a ""tourist rip-off"" or ""tourist trap"".141 142Finally, send me these 3 options, including the city, round-trip flight price, and the recommended restaurant."
143task-19fcf3,"I have a $100,000 down payment and want to buy a condo in Nashville's The Gulch neighborhood for short-term rental investment.144 145First, go to Redfin and search for 'Condo' in The Gulch area, with a price under $600,000. To ensure cash flow, please use the filter to find a property with relatively low HOA fees (preferably under $500/month). Record its total price and estimated monthly payment.146 147Next, go to Airbnb, locate the same The Gulch area on the map, and search for accommodation prices for a weekend next month (e.g., Friday to Sunday). Set the filters to 'Entire place' and a rating of 4.8 or higher (or 'Guest Favorite' tag). Find 3 comparable listings with the same number of bedrooms as the Redfin property, and calculate their average nightly price.148 149Finally, help me calculate: If the property is rented out for 20 days per month at this average nightly rate, can the income cover the Redfin property's monthly payment?"
150task-1a7a24,"I am producing an ""Annual Indie Game Review"" video and need to collect materials for three games: Pacific Drive, Balatro, and Animal Well.  For each game, please perform the following steps in order:151 1521.  First, on YouTube, search for the game's ""Official Launch Trailer"". Filter for the video with the highest views that is published by the official developer or publisher channel, and ensure it supports 4K or 1080p resolution.1532.  Next, go to Metacritic and search for the game's PC version page. Record its User Score. Then, find a user review with a score of 10 and extract a core statement from it.1543.  Finally, on Twitch, search for the game's category. Navigate to the ""Clips"" section, set the filter to ""All Time"", and identify the clip with the highest view count.155 156Please output the following for each of the three games: YouTube Trailer Title, Publishing Channel, Metacritic User Score, Selected Review Summary, Twitch Most Viewed Clip Title, View Count, Streamer Name, and all relevant detail page links."
157task-1b8b3c,"I'm planning to go camping at Grand Canyon Village tomorrow with my Subaru Outback. First, please go to AccuWeather and check the hourly weather forecast for Grand Canyon Village for tomorrow. I'm particularly interested in the temperature and wind speed at 2 PM to see if it's suitable for pitching a tent. Next, to ensure the car can handle the terrain and my loading needs, go to Car and Driver and find a detailed review of the Subaru Outback. I need you to delve into its 'Specs' page to find out its exact Ground Clearance in inches, and its Max Cargo Volume in cubic feet with the rear seats folded down. Also, please check what the first item listed under 'Highs' (pros) by the editors in the review summary is."
158task-1c3247,"I’m planning a short getaway on a weekend next month (check in on Friday, check out on Sunday), but I have fairly severe allergies and am deciding between Austin, TX and Seattle, WA. First, check AccuWeather’s **Health & Activities** forecast for both cities for that travel weekend and compare the **Mold** or **Pollen** risk levels. If the forecast for that weekend is not yet available, use the nearest available date within the next two weeks and clearly note the date used. For extra assurance, then check EPA AirNow for the current **AQI** values in both cities. Based on this data, choose the city with more allergy-friendly air conditions.159 160After selecting the city, go to TripAdvisor and find 3 hotels in that city. Requirements: stay dates must match my trip window (Friday to Sunday of next month), nightly price must be between **$200 and $450**, hotels must be **4-star or above**, and they must include **Air Conditioning** to reduce allergy risk from open windows. Finally, sort by **Traveler Ranked** and list the top 3 hotels.161 162Output required: mold/pollen levels for both cities, AQI values for both cities, the final selected city, and for each recommended hotel: name, price, star rating, whether air conditioning is included, and the detail page link."
163task-1d25d4,"I am currently in Portland, OR, but I suffer from allergic rhinitis and am debating whether or not to visit the Portland Japanese Garden. Please help me make a decision.164 165First, go to AccuWeather and check the allergen situation for Portland for the next three days. Do not just look at the general weather; I need the specific levels (e.g., Low/High) for 'Tree Pollen' and 'Mold'. Please also try to find the dedicated Health or Allergy page on their site.166 167Then, go to Google Maps, search for Portland Japanese Garden, click into the reviews section, and scroll through recent comments to see if any visitors have complained about pollen allergies or mentioned that plants are currently in full bloom.168 169Synthesize the information from both sources and tell me which of the next three days would be most suitable for a visit, or if I should just avoid going altogether."
170task-1e1059,"Next Friday, I’m planning a family vacation to Miami and need two contingency plans based on the weather.171 172First, check AccuWeather for the **daytime precipitation probability** in Miami for the travel day.  173- If the precipitation probability is **above 40%**, go to TripAdvisor and find **3 indoor things to do** that are **Good for Kids**, sorted by **rating (highest to lowest)**.  174- If the precipitation probability is **40% or below**, find **3 outdoor things to do** that are **Good for Kids**, also sorted by **rating (highest to lowest)**.  175 176If AccuWeather cannot provide the precipitation probability for that travel day, use the **daytime precipitation probability for the nearest available date** and apply the same threshold logic.177 178Return the precipitation probability value for the day (or nearest available date), along with each recommended activity’s **name, rating, review count, and TripAdvisor detail page link**."
179task-1e71cc,"I want to plan my Python data engineering study for 2026.180First, go to Amazon and search for ""Python Data Engineering"" paperback books. Use the filtering options to identify 3 candidate books that have a rating of 4 stars or higher, are priced between $30 and $60, and were published in 2024 or 2025.181Next, go to Goodreads, look up these 3 books, and select the one with the highest ""Rating Count"" as the final choice.182Finally, extract core technical keywords mentioned in the title of the ultimately selected book (e.g., PySpark, Kafka, Airflow). Then, go to StackOverflow, search for the tag corresponding to this technology, navigate to that tag's page, sort by ""Votes,"" and find the question with the highest number of votes.183 184Please output: the titles/prices/publication years of the 3 candidate books, the Goodreads rating count of the finally selected book, the extracted keyword, and the title and link of the highest-voted StackOverflow question."
185task-1ea4cc,"I need to reproduce a **Music Source Separation** algorithm and want to find several reliable open-source implementations.186 187First, search arXiv for papers on Music Source Separation from the **last six months**, and pick **three papers** that appear to be relatively influential. Then, for each of these three papers, search GitHub for the corresponding official repository.188 189For each repository found, focus on checking:190 1911. Whether it is an **Official Implementation**;  1922. Whether there have been any code updates (**commits**) in the **last two months**;  1933. Whether it is a complete project or just a skeleton (e.g., check whether the file tree includes actual `src` or training code, not just README files).194 195If no official implementation is found for a paper, look for a relevant community reproduction repository instead (prefer ones with higher star counts) and clearly label it as **“Unofficial.”**196 197Finally, compile a checklist that includes the paper title, the corresponding GitHub repository link, and conclusions for each of the checks above."
198task-22fc0a,"I’m going to San Francisco this coming Friday to see **“The Racket”** stand-up show at **Cobb’s Comedy Club**, which starts at **8:00 PM**. Please help me plan what to do after the show: first, confirm approximately what time the show will end, then find **3 restaurants on Yelp** that are within a **10-minute walk** and still open at that post-show time. Prefer places rated **4.0 stars or higher**, and **no fast food**.199 200Please provide the **name**, **rating**, **exact walking time**, and **closing time** for each of the 3 places. Also, please check which one has Yelp reviews that mention **“late-night snack”** or **“late night”** most frequently."
201task-25f34a,"A family member of mine was recently diagnosed with hypertension. The doctor recommended a low-sodium diet, but he is allergic to shellfish and peanuts. I want to create a one-week dinner plan that meets hypertension dietary guidelines while avoiding allergenic ingredients.202 203First, check Mayo Clinic for dietary recommendations for people with hypertension. Specifically record the daily sodium intake target relevant to blood pressure management. If the page does not provide a “hypertension-specific” value, record the sodium recommendation given for adults/cardiovascular health and clearly note the applicable population. Also capture the recommended food categories.204 205Then go to WebMD to confirm possible cross-reactive foods for shellfish and peanut allergies (e.g., whether other seafood or legumes are safe).206 207Next, search AllRecipes for low-sodium dinner recipes. Find 5 recipes rated above 4 stars with no more than 600 mg sodium per serving. Review each recipe’s full ingredient list carefully, exclude any recipe containing shellfish, peanuts, or identified cross-reactive ingredients, and keep 3 final safe recipes.208 209For the core protein ingredients in these 3 recipes (e.g., chicken, beef), check the EWG database for food safety ratings, ensuring there are no pesticide-residue or additive-risk concerns.210 211Output:212- Summary of Mayo Clinic hypertension diet guidance (daily sodium limit, recommended food types, and the applicable population for the sodium value) + page link  213- WebMD-confirmed list of cross-reactive allergy foods + page link  214- For 3 qualifying recipes: recipe name, link, sodium per serving, full ingredient list, and user rating  215- EWG safety rating + link for each recipe’s core protein ingredient"
216task-26914d,"I've recently developed a keen interest in mythological motifs on ancient Greek pottery and would like to study them systematically.217 218First, please go to the Getty Museum website and find three examples of 5th-4th century BCE Greek red-figure pottery. Ensure that the description for each piece explicitly mentions specific Greek mythological figures (e.g., Athena, Heracles, Achilles, etc.). Record the name of the pottery, the mythological figure(s) depicted, the date, and the link to the Getty detail page.219 220Next, use the names of these three mythological figures to search on Google Scholar. For each figure, find two English academic papers published since 2020 with more than 10 citations, whose themes are related to Greek pottery art or mythological iconography.221 222Finally, go to YouTube and search for explanatory videos using the English name of each mythological figure. Filter for educational content between 10-30 minutes in length with over 5,000 views. Find the single most relevant video for each figure, and expand its description to confirm whether it mentions ancient Greek art or pottery.223 224**Output:**225*   **For each piece of pottery:** name, date, mythological figure(s) depicted, Getty link.226*   **For each paper:** title, author(s), publication year, citation count, Scholar link.227*   **For each video:** title, duration, view count, channel name, a summary of the video description (highlighting whether it mentions pottery or art), and YouTube link."
228task-273c1b,"I’ve recently been exploring fine-tuning techniques for large language models, and I heard there’s a classic paper on **“Llama 2”** fine-tuning (the title likely includes “Llama 2” and “Open Foundation”). Please help me find this paper on arXiv.  229 230After finding it, I want to check whether there are any official or highly popular community reproductions on Hugging Face. First, extract the paper’s **Code link** or **project homepage** from arXiv (if no usable code/project link is provided on the page, then use the paper title or author names as search clues), and then search for the corresponding model on Hugging Face.  231 232Find the version with the highest download count, and check whether its Model Card explicitly cites or mentions which dataset(s) were used. Also confirm whether this model is officially released by **meta-llama**.  233 234Finally, tell me the **model name**, **whether it is official**, and the **dataset name(s) mentioned**."
235task-2a47cb,"I’m arranging an interstate consultation trip with gastroenterology specialists for my father, and I need to screen doctors and plan the itinerary. First, go to health.usnews.com and find one **Gastroenterologist** at **Mayo Clinic in Rochester, Minnesota** and one at **Cleveland Clinic in Ohio**. Both doctors must meet two strict criteria: **more than 20 years of experience** and a **perfect 5-star patient rating** (if no doctor at that location has a full 5 stars, relax the rating requirement to **4.5 stars or higher** and clearly note the reason for the relaxation).  236 237After identifying the doctors, go to maps.google.com and calculate the estimated **air travel time** from **Mayo Clinic (Rochester)** to **Cleveland Clinic**.  238 239Finally, return to the map view for **Mayo Clinic (Rochester)** and find a hotel. Prioritize hotels that meet both conditions: **within 1 mile walking distance** and **rated above 4.5**. If no hotel satisfies both at the same time, keep the rating threshold at **≥4.5** and relax walking distance to **within 1.5 miles**, and explicitly state this relaxation in the output.  240 241Output the following: both doctors’ names, years of experience, ratings, flight travel time between the two locations, and the selected hotel’s name, walking distance, rating, and all detail-page links."
242task-2a9f91,"I just received a software engineer offer from a tech company in the San Francisco Bay Area, with an annual salary of $120,000, and now I need to find housing. The company is Salesforce. First, I want to confirm the exact address of their San Francisco office (preferably from the company’s LinkedIn page; if LinkedIn does not show the full address, use Salesforce’s official office locations page or the official Google Maps location page to verify it, and cite the source).243 244Then help me find apartments near that address with a budget of no more than $2,500/month, ideally with a driving commute of no more than 30 minutes, and it must be a 1-bedroom unit. My top priorities are an in-unit washer and dryer and a parking space, since I plan to drive to work.245 246Please try to find 3–5 listings that meet these criteria. If there are fewer than 3 under the strict requirements, keep the existing results and state the count—do not relax any of my filters. For each listing, calculate the total move-in cost, including deposit and parking fees, and estimate the monthly commute time.247 248Finally, summarize which options offer the best value for money so I can plan in-person tours."
249task-2bf0e1,"Lately my commuting time has been pretty fragmented, so I’m looking for a few highly rated, short-play RPGs suitable for Steam Deck.  250First, go to Metacritic and filter for RPG titles with a Metascore above 85. Platform can be either PC or Switch. Sort by user rating (if that sort option is unavailable, use the closest user-related sort available on the site, or keep the default sort and continue). Check multiple pages rather than only the first popular page.251 252Then pick 3 titles you think are good candidates, and look up their **Main Story** playtime on HowLongToBeat. Prioritize games with a main story of **20 hours or less**. If there are fewer than 3 titles strictly under 20 hours, keep the **Metascore ≥ 85** requirement unchanged and fill the list to 3 from the candidate pool, clearly marking which ones exceed 20 hours.253 254Finally, go to Steam and confirm whether these 3 games are Steam Deck verified (shown with the green **Deck Verified** icon).255 256Please output the following for the 3 games: **title, Metascore, HLTB Main Story length, and Steam Deck compatibility status**."
257task-2d3de9,"I just watched 'Dune 2' and want to delve deeper into the story behind the film.258 259First, go to IMDb, find the movie, and note down the director's name, the original author from the 'Writers' section, and the title of the original book.260 261Then, search for the original book title on Goodreads. Once found, provide its rating, number of reviews, and page count. If it's part of a series, list the titles of other books in the same series.262 263Next, go to YouTube and search for 'Dune 2 behind the scenes' or 'Dune 2 making of'. Filter the results to videos longer than 15 minutes, sort them by view count in descending order, and identify the top 3 videos that are behind-the-scenes/making-of specials.264 265Finally, visit the official Criterion Collection website and search for the director's name. Check if any of his other works are featured. If so, list the title of the work, available formats (DVD/Blu-ray/4K), and a brief description of the special features.266 267**Output:**268Movie Director's Name, Original Book Title, Original Author, Goodreads Rating, Number of Reviews, Page Count, Other Series Titles (if applicable), Goodreads Book Link, Titles of the top 3 YouTube videos, View Counts, Durations, Video Links, Titles of other director's works featured on Criterion, Formats, Brief Description of Special Features, Criterion Collection Detail Page Links."
269task-2f4ebf,"I’m planning to drive my 2024 Tesla Model Y Long Range on a weekend next month from Griffith Observatory in Los Angeles to the Bellagio Hotel in Las Vegas. I don’t trust the official EPA range, so first go to Car and Driver and find the real-world “75-mph highway range” test result for this model. If the exact model’s 75-mph test data is unavailable, then find the tested 75-mph range for a same-year or nearby-year Model Y Long Range on Car and Driver, and clearly note any model/year differences. Use 80% of that tested range as my safe maximum distance for the first driving leg.270 271Next, plan the route in Google Maps and use the “Search along route” feature to identify 3 Supercharger stations located between 150 miles from the origin and that safe maximum distance (if fewer than 3 are available, record all that can be found and note the count).272 273Finally, for each of those 3 charging stations, go to Yelp and find the highest-rated restaurant within 0.2 miles of the station’s address (rating must be 4.0+ and price level must be $$ or lower). If no qualifying option exists within 0.2 miles, expand the radius to 0.5 miles and explicitly note that this relaxation was applied.274 275Output: the Car and Driver tested 75-mph range, the calculated safe maximum distance, and the C&D article link; the names of 3 candidate charging stations and their distances (in miles) from the origin; and for each station, the recommended restaurant name, Yelp rating, price tier, and Yelp detail page link."
276task-2fd0d4,"As a film noir enthusiast, I want to collect three classic films released between 1940 and 1959 that are highly acclaimed.277 278First, on Metacritic, filter for film noir titles released between 1940 and 1959 with a Metascore of 85 or higher. Sort them in descending order by score and select the top three.279 280Next, visit the official Criterion website to confirm if Blu-ray versions (non-4K, non-DVD) of these three films are available for purchase. Check the technical specifications for a ""4K digital restoration"" note and record the official Criterion website price.281 282Finally, search on eBay for used Blu-ray discs of these three films. Requirements: condition ""Very Good"" or better, ""Buy It Now"" format only, and the total price (including shipping) must be lower than the official Criterion website price.283 284Output the following for each film: Title, Metascore, Criterion official Blu-ray price, whether it has a ""4K digital restoration"" note, and the lowest eligible eBay item price and link."
285task-30b78c,"I am conducting research on Neural Radiance Fields (NeRF), and my advisor asked me to compile the latest NeRF-related work from CVPR 2024. First, go to OpenReview and search for papers in CVPR 2024 that contain the keywords “NeRF” or “Neural Radiance Fields.” Filter for papers with acceptance status **Oral** or **Spotlight**, and collect up to 5 papers (if fewer than 5 are available, record the actual number). For each paper, record the title, authors, and OpenReview link.  286 287Then, use each paper’s title and first author name to search arXiv for the corresponding preprint. After finding a match, record the arXiv ID and PDF link.  288 289Finally, go to Bilibili and search for explanatory videos using core keywords from each paper title (e.g., “神经辐射场” or key English title terms). Prefer one Chinese explanatory video per paper with duration over 10 minutes and more than 10,000 views. If no video for a paper satisfies both conditions, keep the core requirement of being a Chinese explanation video, select the most relevant one with relatively higher views or longer duration, and note which condition(s) are not met.  290 291For each paper, output: paper title, first author, acceptance type (Oral/Spotlight), OpenReview paper page link, arXiv ID, arXiv PDF link, Bilibili video title, video duration, view count, video link, and whether the video meets the criteria “duration ≥ 10 minutes and views ≥ 10,000.”"
292task-30e4ca,"I’m preparing a submission to CVPR and want a comprehensive understanding of the conference. First, please check DBLP to analyze the topic distribution of CVPR papers from the past three years, with a focus on identifying major trending areas. Then, on arXiv, find several highly influential papers in those trending areas (judged using both visible paper-page signals and external citation-metric pages), and review their writing style and experimental design. Next, go to GitHub to locate the corresponding open-source code repositories for those papers, and evaluate their star counts and code completeness to understand current experimental standards. Finally, search Bilibili for CVPR-related peer-review experience sharing or conference talk videos with high view counts and durations over 20 minutes, and summarize the key reviewing points that are most frequently emphasized.  293 294At the end, provide me with a structured summary of: what directions this conference currently values most, what experiments the reference papers typically include, expected code quality standards, and what reviewers pay special attention to."
295task-315905,"Our team is evaluating environmental compliance risk in Harris County, Texas. Please use the EPA ECHO system and prioritize list-based results (do not rely on the map view) to identify one chemical or manufacturing facility in the county that is currently in **Significant Non-Compliance (SNC)** status. First, open the facility ranked at the top of the list, review its detailed report, and determine the specific chemical pollutant for which it is in violation (e.g., benzene, lead, particulate matter).296 297If the map page appears blank, flickers, or fails to load reliably, switch to an accessible list/table results page and continue the screening there. If the first facility’s detail page does not clearly name a pollutant, move on to the next facility, and continue until you find one with an explicit pollutant record.298 299After obtaining the pollutant name, conduct a technical search in the USPTO patent database. Find **two U.S. patents granted after 2021** that are specifically aimed at treating or removing that pollutant. Finally, report the facility name, the violating pollutant, and the titles of the two patents you found."
300task-328b3d,"I am planning a ""Game vs. Reality"" pilgrimage video for ""Like a Dragon: Infinite Wealth,"" focusing on Honolulu, Hawaii.301 302First, navigate to IGN and search for ""Real Life Locations"" or a similar guide for the game. Identify a list that maps in-game establishments/landmarks to their real-world counterparts.303 304From this list, select four real-world commercial establishments (restaurants, shops, or bars, excluding generic parks or streets). Each selected location must have a Google Maps rating of 4.0 stars or higher to ensure their quality and appeal.305 306For each of these four selected locations, search for their real-world names on Google Maps. Verify their specific ""price level"" (e.g., $, $$) and their current ""operating status"" (e.g., Open or Closed).307 308Output the following for each location: In-game Name, Real-world Counterpart Name, Real-world Address, Google Maps Rating, Price Level, Operating Status, and the respective links to the IGN guide page and the Google Maps detail page."
309task-328b4b,"I’ve just finished learning Python fundamentals and want to switch into data analysis, but I’m not sure whether pandas or PyTorch is currently more in demand.  310First, go to Stack Overflow and open the two tag pages ([pandas] and [pytorch]). Record each tag’s **“Total questions”** count. Then, under each tag, sort by **Newest** and count the number of new questions posted in the **past 24 hours** as today’s activity level. Compare the two and identify which library is more active.311 312Next, for the winning library, go to Udemy and find **3 advanced courses** that meet all of the following criteria:  313- Rating above **4.5**  314- **English-language** instruction  315- Duration over **17 hours**  316- Level set to **Expert** or **Intermediate**317 318If fewer than 3 courses meet all criteria, keep the rating, language, and duration requirements unchanged, first relax the Level filter to **All Levels**, and clearly mark which courses were added after relaxing this condition.319 320Output required:  321- For Stack Overflow: total question count and past-24-hour question count for both libraries, plus the name of the winning library  322- For Udemy: the 3 course titles, instructors, ratings, durations, prices, and course links"
323task-32aa99,"I want to refresh my bedroom in a Mid-Century Modern style this month. Please create a soft-furnishing shopping list with **three items**, with a **total budget of no more than $400**.324 325First, go to **IKEA**, search for **“Chest of drawers,”** filter for **Brown/Wood** finishes, and sort by **price from low to high**. Choose one **3-drawer or 4-drawer** chest that is currently marked **“In stock”** and priced **under $200**.326 327Next, go to **Etsy** and find a **“Bauhaus exhibition”** art poster that matches this style. It must be a **physical item** (not a digital download), from a **Star Seller** shop, with a size of **11x14 inches or larger**.328 329Finally, go to **Target** and find a **table lamp** with a **Brass/Gold** finish, rated **4 stars or higher**, and available for **Shipping**.330 331For each item, provide: **product name, price, key verification info** (IKEA stock status, Etsy shop badge, Target rating), **size/specs**, and the **product detail page link**."
332task-32d292,"I’m preparing for a Google interview and want to understand the latest research trends in the **Machine Learning** direction at their Mountain View headquarters.333 334First, search LinkedIn for a **Research Scientist** role at Google in **Mountain View**, and find one that was posted recently (**Past Month**) or otherwise appears relatively new. Carefully review the job description and extract one core technical keyword (such as a specific large-model technique or algorithm name). If you encounter a LinkedIn login requirement or cannot view full job details, record that access limitation and extract the keyword from whatever LinkedIn-visible job summary is available; if the summary is insufficient, then extract one keyword from visible skills/requirements on the same page.335 336After obtaining the keyword, go to Google Scholar and check Google’s publications in that area. Focus on papers from **2023 onward** and identify the top 3 by citation count (**Cited by**). For each of the three papers, provide the title, first author, and citation count, and also verify whether the author list appears to include Google-affiliated researchers (typically indicated by institution tags/abbreviations next to names). If Google Scholar triggers unusual traffic checks/CAPTCHA or cannot be accessed reliably, switch to Semantic Scholar and use the same keyword and year filter (**2023 onward**), then select the top 3 by citations and output the same fields (title, first author, citation count), plus the basis for Google-affiliation judgment (author institution or institution info on the paper page)."
337task-3559ff,"I’m planning to attend some art-related activities in Los Angeles this coming weekend and want an itinerary that feels full but not rushed. First, check the Getty website for exhibitions or events happening this weekend (lectures, workshops, guided tours, etc.), choose one that looks interesting, and record it. If there are no suitable options this weekend, then find the nearest upcoming Getty event within the next two weeks and note the reason for this adjustment.338 339Then go to Eventbrite and search for art-related events in Los Angeles for this Saturday and Sunday. Under the Art, Museums, or Culture categories, find 2–3 well-rated events. If the platform does not provide reliable rating data, use visible indicators such as registration popularity, follower/interest counts, or comments as evidence of being “well-rated,” and specify the basis in the output.340 341Arrange the Getty event and the Eventbrite events in chronological order, and use Google Maps to calculate driving time between each pair of consecutive event locations.342 343Help me filter for one feasible combination with these constraints:344- There must be sufficient time between every two events:  345  previous event end time + travel time + at least 30 minutes buffer ≤ next event start time.346- Keep 3–4 events in the final plan with no time conflicts.347- If fewer than 3 events can satisfy the constraints, keep all feasible events and explain why the count is insufficient.348 349For each event, output:350- Event name351- Start time352- End time (or duration)353- Venue address354- Brief description355- Event page link356 357For each adjacent event pair, output:358- Google Maps route link359- Driving distance360- Estimated travel time"
361task-360097,"I’m a beauty influencer looking for a few recently trending foundations on Sephora to recommend to my followers. First, search for “foundation” on Sephora, sort by best sellers, and identify 5 popular foundations with ratings above 4.5 and prices between $30 and $60. Record the brand name, exact product name, shade range (at least how many shades are available), price, and rating.362 363Then, search on Target for foundations from the same brands as these 5 Sephora products, prioritizing the exact same item when possible (matching name/line preferred), and check whether Target offers a lower price. Record Target’s price and stock status.364 365Finally, visit each of the 5 brands’ official websites (e.g., Fenty Beauty, NARS, etc.). Prioritize locating the product page corresponding to the same item selected on Sephora, and check whether the official site has bundle deals or promotions not available on Sephora/Target (such as a free mini with full-size purchase, member-exclusive pricing, etc.). Capture the official product page URL and promotion details.366 367If an exact match for any item cannot be found on Target or the brand’s official site, keep that item and mark the corresponding field as “Not found,” then continue with the remaining items.368 369Output fields for each foundation:370- Brand name  371- Full product name  372- Sephora price  373- Sephora rating  374- Number of shades  375- Target price  376- Target stock status (In stock / Out of stock / Not found)  377- Brand official website price  378- Official website promotion details (if any)  379- Sephora product URL  380- Target product URL  381- Brand official website product URL"
382task-391330,"I want to go to an event in Seattle within the next two weeks (preferably next Saturday) related to “AI” or “Technology.” First, search on Eventbrite and pick one event that fits the timing and appears relatively formal, then note the exact venue address. Next, go to TripAdvisor and find a hotel very close to that venue, with a requirement of at least 4 stars and a rating of 4.0 or higher. Finally, to confirm convenience, use Google Maps to plan a walking route from the hotel to the event venue, prioritizing options with a walking time of 15 minutes or less. If no option within 15 minutes meets the hotel criteria, keep the hotel criteria unchanged and choose the feasible hotel with the shortest walking time, then state the actual walking time in the output. Output the event name, hotel name, and the walking minutes shown on Google Maps."
383task-3a846e,"I recently found a house in the distant suburbs of Houston that I'm interested in, located at 20306 Telge Rd, Tomball, TX 77377. Although the price is attractive, I'm concerned about high future commuting costs. I currently drive a 2022 Toyota RAV4, and I work in downtown Houston (using JPMorgan Chase Tower as the reference point).384 385Please help me calculate the commuting cost:3861. First, check on Zillow to confirm if this house is still for sale. While you're there, please also find its build year and listing price.3872. Use Google Maps to calculate how long it takes to drive from this house to JPMorgan Chase Tower at 8 AM on a Monday morning, and what the approximate one-way distance in miles is.3883. Go to Car and Driver to find the actual highway MPG (miles per gallon) for my car (2022 Toyota RAV4) in their reviews.3894. Finally, help me calculate the total: Assuming one round trip per day, 22 working days per month, and a gas price of $3 per gallon, how much would I likely spend on commuting fuel costs each month if I don't change my car?"
390task-3a8e81,"I’m self-studying Prof. Patrick Winston’s *Artificial Intelligence* (6.034) on MIT OCW. The course is classic, but I want to connect it with the latest research.391 392First, search for this course on OCW, and make sure to check the **“Video Lectures”** filter in the left sidebar so we use a version that includes videos. After opening the course page, check the **Calendar** or **Syllabus** and find the topic title corresponding to **“Lecture 12.”**393 394Then, go to arXiv and use **Advanced Search**. Under the **Computer Science** category, search for that topic. Crucially, you must use the date selector to restrict the time range to the **most recent 24 months up to today**.395 396From the results, find the **first paper** whose **Comments** or **Abstract** explicitly mentions **“GitHub”** or **“Code available,”** and tell me its **title** and **arXiv ID**.397 398If there is no result under those conditions, keep the same topic and Computer Science category, widen the time range to the **most recent 36 months up to today**, search again, and explicitly note in the output that the relaxed time window was used."
399task-3bf7cd,"I've recently been watching 'Inception' and was thoroughly impressed by its non-linear narrative and multi-layered dream structure. I'd like to systematically learn about such complex narrative techniques and screenwriting skills.400 401First, please go to the IMDb detail page for this movie. Specifically, focus on the technique-related keywords discussed by professional film critics and screenwriting enthusiasts in the comments section. Identify the screenwriting techniques mentioned, such as narrative structure, timeline management, and character arcs.402 403Then, using these keywords, search Coursera and edX for relevant film production or screenwriting courses. Ideally, these courses should have a rating of 4.5 or higher and be backed by a university or reputable institution. Please find 3-5 such courses, noting down the course title, platform, institution, rating, and main content focus.404 405Finally, on Amazon, search for screenwriting textbooks using these technique keywords. Find books with a rating of 4 stars or higher. Please select the 5 most relevant ones, examining the book cover, title, author, rating, and reader reviews to confirm they indeed cover these techniques.406 407Lastly, please compile a complete learning roadmap for me: What core techniques did the film use → Which courses systematically teach these → Which books can provide in-depth learning."
408task-3d05a3,"I just listened to Kanye West's 'Bound 2' and particularly loved the high-pitched vocal sample in it.409Please go to YouTube and find this song. Check the comments section to see if anyone mentions the name of the original song from which the sample was taken (remember to scroll through comments or sort by top comments).410Once the original song's name is identified, search for it on Genius.com. On the lyrics page, examine the annotations (by clicking on the highlighted lyrics) to confirm which specific lyric line was sampled in 'Bound 2'.411Finally, I want to purchase a vinyl record of this original song for my collection. Go to Discogs.com, search for the album containing this original song, filter the format by 'Vinyl', and find an available item with a 'Media Condition' rating of 'Near Mint' or 'Mint' and priced under $50. Provide me with the purchase link."
412task-3e6e24,"I am planning to fly from San Francisco to Austin this coming Friday to Sunday to attend a music festival. Please help me plan a complete weekend itinerary:413 414First, go to Eventbrite and search for music festival events in Austin for this coming Friday to Sunday. Find a large electronic or rock music festival and note its start time, end time, and specific venue address.415 416Then, use Google Flights to search for round-trip flights from San Francisco to Austin. The outbound flight must arrive in Austin at least 4 hours before the festival begins. The return flight should depart on the morning following the festival's end. Only consider direct flights, with a price under $500.417 418Next, go to Airbnb to find accommodation near the festival venue. Check-in on this coming Friday and check-out on the following Monday. Use Google Maps to verify that the walking distance from the accommodation address to the festival venue is indeed within 2 miles. Filter for listings with a rating of 4.5 or higher and a price under $150 per night.419 420Finally, use Yelp to search for restaurants near the accommodation address. Find three restaurants with a rating of 4 stars or higher, a price level of $$ or below, and that are currently open.421 422Output: Festival name, start time, end time, venue address, Eventbrite event page link; Outbound flight number, departure time, arrival time, price; Return flight number, departure time, arrival time, price, Google Flights booking link; Accommodation name, address, rating, price per night, walking distance to festival venue, Airbnb listing link, Google Maps distance verification screenshot link; Names, addresses, ratings, price levels, operating hours, and Yelp page links for the three restaurants."
423task-3fbc94,"Next Saturday, we are taking elderly family members and children to visit the Getty Villa in Los Angeles (please note: Villa, not Center). We expect to finish around 6 PM and need to find a place for dinner nearby. On TripAdvisor, starting from the Getty Villa, find 3 restaurants within a 10-minute drive, with a rating of 4 stars or higher. Please consider the preferences of our elderly family members, prioritizing Italian or seafood cuisine. Additionally, check reviews for mentions of ""family-friendly"" or ""quiet."" After selecting these 3 restaurants, check OpenTable for availability for 6 people between 6 PM and 7 PM next Saturday evening. Finally, list the names of these 3 restaurants, their ratings, positive aspects mentioned in reviews, and the specific available reservation times on OpenTable."
424task-3fc3ea,"I’m planning to visit the Getty Center in Los Angeles this weekend (Saturday) to see an exhibition. First, check the official website to see whether there are any exhibitions in the **Current Exhibitions** section related to **ancient history** or **medieval art**, and confirm that your selected exhibition is open to the public on the planned travel day. After choosing the most interesting one, record its name.425 426Then, since I don’t want to spend the whole day only at the museum, search Eventbrite for Saturday afternoon or evening events near the Getty Center (in the Brentwood or Westwood area). Find one event in either the **“Music”** or **“Performing & Visual Arts”** category, preferably **free** or priced at **$50 or less**, and note its time and location.427 428Finally, use Google Maps to plan the route: assume I depart from Santa Monica at 10:00 AM, go to the Getty Center first, and then attend that event. Along the way, use **Search along the route** to find an Italian restaurant with a rating of **4.5 or above** for dinner. If there is no Italian restaurant rated 4.5+ along the route, choose the highest-rated Italian restaurant available along the route and specify its rating.429 430Please compile a complete itinerary schedule and the total travel time for the full route."
431task-426392,"I’m working on a climate data monitoring project and want to regularly and automatically pull the latest climate data from NOAA (National Oceanic and Atmospheric Administration). Please help me find a climate-related dataset on data.gov that has an API, has been updated within the last 3 months, and is available in CSV format. Check the API documentation for the endpoint URL and authentication method.432 433Then, search GitHub to see whether anyone has written Python scripts for calling this type of federal data API. Find a project with as many stars as possible, still maintained within the past year (prefer non-archived), and review how the code handles API authentication and data parsing, including whether it implements pagination or rate-limiting logic.434 435If no dataset on data.gov meets all of these criteria (“updated within the last 3 months + CSV + NOAA + API”), relax the update window to the last 6 months first. If there is still no match, keep the NOAA and API requirements and record the actual available format.436 437Finally, provide a structured summary of the dataset’s API endpoint, required authentication parameters, and the key code logic from the GitHub project, and clearly indicate whether each condition is satisfied."
438task-42c042,"I'd like to purchase three highly replayable and well-regarded games during the current Steam sale. My budget is limited, with each game not exceeding 50 USD.439 440First, navigate to Steam's 'Specials' or 'Deals' section. Find several discounted games that are tagged with both 'RPG' and 'Indie'. Make sure to browse through multiple pages, not just the first one. Keep a record of these candidate games' names and their discounted prices.441 442Next, visit HowLongToBeat.com to check the 'Main Story' duration for each candidate game. Eliminate any 'quick-play' games with a Main Story duration shorter than 15 hours. Subsequently, go to Metacritic.com for the remaining games. Check if their Metascore is 80 or above, or if they have received the 'Must-Play' golden badge.443 444Finally, present the top three 'high-value' games that meet these criteria. For each selected game, provide its name, Steam discounted price, HLTB Main Story duration, Metacritic score, and your recommendation reason. If three perfect matches cannot be found, two suitable recommendations will suffice."
445task-44ddd7,"I recently discovered an electronic musician named 'Ross from Friends' on SoundCloud, and I find his style quite unique. I'd like to learn more about him.446 447Please start by navigating to his official SoundCloud page. Confirm if he is from the United Kingdom (UK) and identify the name of his track with the highest number of plays.448 449Then, search for this artist on the Discogs database. After confirming it's the same individual (by comparing tracks or musical style), please find his earliest physical release (e.g., Vinyl or Cassette). Specify the year of release and the title of this record.450 451Finally, check Wikipedia for an article about him. If an article exists, please provide his real name.452 453Please compile and present these three pieces of information: the SoundCloud track with the highest play count, the year and title of his earliest physical release from Discogs, and his real name (if available)."
454task-462dac,"I am preparing for a live-streamed game series titled '2025's Most Playable Action-Adventure Games'. I need to select 3 high-quality and trending games for in-depth review.455 456First, go to Steam and search for 'Action-Adventure' games. Filter by the 'Action' and 'Adventure' tags, then sort by 'Sales' to identify the top 10 games on the current bestseller list. Record the game names, current prices, and discount information.457 458Next, for each of these 10 games, visit Metacritic to find their Metascore (media rating) and User Score for the PC platform. Retain only games with a Metascore of ≥75 AND a User Score of ≥7.5.459 460Then, go to HowLongToBeat to query the 'Main Story' completion time and 'Completionist' time for these filtered games. Further filter to include only games with a 'Main Story' completion time between 15-40 hours (games that are too short won't provide enough content for a series, and games that are too long might cause viewer drop-off).461 462Finally, search IGN for these remaining games to confirm if a review article exists. Record the IGN score, and check the completeness of the Wiki guide page (specifically, look for detailed quest walkthroughs and collectible location guides).463 464Please output a comparison table for the finally filtered games, including:465*   Game Name466*   Steam Price & Discount467*   Metacritic Media Score468*   Metacritic User Score469*   Main Story Completion Time470*   Completionist Time471*   IGN Score472*   IGN Guide Completeness (Complete/Partial/None)473*   Steam Detail Page Link474*   Metacritic Detail Page Link475*   HowLongToBeat Detail Page Link476*   IGN Review Link477 478From this list, recommend the 3 most suitable games for the live stream series, explaining the reasons based on dimensions such as scores, playtime, trending potential, and guide support."
479task-46b0d7,"I’m planning to go to Miami this weekend for an EDM party, and I need you to arrange a complete itinerary. I’ll be departing from New York and want to find an electronic music event on Saturday night (prioritize the nearest upcoming weekend within the next two weeks; if there is no suitable event that weekend, check the following weekend and state that in the output).480 481After finding the event, book a flight that arrives in Miami at least 4 hours before the event starts (if it’s a connecting flight, the layover must be at least 2 hours). Finally, find an Airbnb near the venue, with a walking time of no more than 15 minutes to the venue.482 483Please provide:484- Event name485- Venue name and address486- Event start time487- Eventbrite link488- Flight number489- Departure and arrival times490- If connecting, include layover city and layover duration491- Time difference between flight arrival and event start492- Airbnb listing link493- Airbnb address494- Walking distance and walking time from Airbnb to venue495- Google Maps walking route link"
496task-46cf17,"I've recently been researching 3D reconstruction techniques related to Neural Radiance Fields (NeRF), and I'd like to reproduce experiments from a representative paper.497 498Please help me search arXiv for papers with titles containing 'Neural Radiance Fields' or 'NeRF', filtering for those published since January 2024 and categorized under Computer Vision (cs.CV). From these, identify 3 seemingly important papers (based on their titles and abstracts), and record their arXiv ID, title, authors, and publication date.499 500Then, search for these 3 papers individually on Semantic Scholar. For each paper, check its TLDR summary, total citation count, and highly influential citation count. Select the paper with the highest total citation count as the target paper.501 502For this target paper, expand to view its complete citation statistics. Switch to the 'Citations' tab, filter for 'Highly Influential' citations, and from these, identify 3 subsequent works that also have a relatively high citation count (indicating that the target paper indeed had a significant impact).503 504Finally, output: the target paper's arXiv ID, title, authors, publication date, arXiv link, TLDR summary, total citation count, highly influential citation count, and Semantic Scholar link; along with the titles, citation counts, and Semantic Scholar links for the 3 highly influential subsequent works."
505task-49e805,"As a Cyberpunk 2077 player, I've encountered an issue where a recent automatic game update has rendered my core mod, 'Cyber Engine Tweaks', inoperable. Could you please assist me in performing a compatibility troubleshooting process:506 5071.  First, navigate to the 'News' section on the Cyberpunk 2077 Steam store page, find, and record the game's latest version number (Patch number).5082.  Then, search GitHub for 'Cyber Engine Tweaks' and proceed to the repository with the most stars.5093.  On the 'Releases' page, check if the latest published Mod version explicitly states support for the Steam game version you just found.5104.  Next, a crucial step: go to the 'Issues' page, sort the issues by 'Recently updated' using the filter, and examine the first few 'Open' issues to determine if there are widespread user reports of crashes for this version.511 512Finally, please provide me with:513*   The current game version number.514*   The recommended Mod version to download.515*   Whether it is advisable to update the Mod at this time, based on the Issues review."
516task-4aed73,"I’m planning a backpacking trip to Thailand, but I take Sertraline and Omeprazole daily and I’m worried about drug interactions.  517Please first check the CDC website to identify which vaccines are listed for travel to Thailand as **“Routine”** and which are **“Recommended.”**  518 519Please pay special attention to the **malaria prevention** guidance and note down the names of all antimalarial medications mentioned on the page.  520 521Then, go to the **Drugs.com Interaction Checker** tool and enter my two daily medications together with all the CDC-mentioned antimalarial drugs to evaluate whether there are any serious interactions.  522 523Finally, tell me which antimalarial option has the **lowest side-effect risk** for me (i.e., the fewest interactions or the mildest interaction severity). If there is a tie, list all tied options and explain the interaction severity basis for each."
524task-4cd281,"I want to systematically learn MLOps (Machine Learning Operations) and plan to earn a certificate on Coursera. Please help me find relevant courses, but I have two strict requirements: the difficulty level must be **""Intermediate""**, and the duration must be **""1–3 Months.""**  525 526To ensure quality, please prioritize sorting by **Highest Rated**. If that sort option is unavailable on the current page or only visible after login, then, among the results that already match the two filters above, choose the course with the highest rating from the top few visible results and state the basis for your selection. After selecting a course, note the instructor’s name.  527 528Since I’m concerned the teaching style might be too dry, I’d like to verify it. Please search YouTube for the instructor’s name and find a publicly available talk or long-form teaching video with the **highest view count** and a duration of **more than 20 minutes** (no Shorts or brief promo clips).  529 530Finally, provide the **course title**, **instructor name**, and the **YouTube video title and link**. If you encounter login/paywall restrictions on Coursera that prevent viewing full course details, record where the restriction occurs and continue the task using all visible information."
531task-4d371c,"I want to undertake a systematic study of the 'Llama 3' model.532 533First, please search for 'Llama 3' on HuggingFace and identify the model version with the highest download count that is tagged for 'text-generation'. Then, navigate to its Model Card. Specifically, examine the 'Uses' section for applicable scenarios and note the Deep Learning Framework (e.g., PyTorch or TensorFlow) indicated in the sidebar.534 535Once the framework and task type are confirmed, proceed to Coursera to find three relevant Specializations. These Specializations must include either 'Generative AI' or 'Natural Language Processing' as keywords, be rated as 'Intermediate' difficulty, and have a minimum rating of 4.5 stars. Please browse through multiple pages to check for courses offered by reputable institutions such as DeepLearning.AI or Stanford.536 537Finally, provide me with the name of the identified model and the names, institutions, and ratings for the three selected courses."
538task-4d3b21,"I am a Manchester United fan, planning a trip to Newcastle for an away game.539Please help me plan my itinerary for December 2025.540First, go to ESPN to find the Manchester United away match against Newcastle United, and confirm the exact date and kick-off time.541Next, go to TripAdvisor to find accommodation for the night of the match. Filter for 3 hotels in Newcastle upon Tyne with a rating of 4 stars or higher, and explicitly stating 'breakfast included'.542Finally, to ensure easy access to the stadium, use Google Maps to calculate the walking route from each of these 3 hotels to St. James' Park. Only retain hotels with a walking time of 15 minutes or less (if all exceed 15 minutes, retain the one with the shortest walking time).543Output the match date and time, the names of the 3 hotels, their TripAdvisor ratings, price per night, whether breakfast is included, and the specific walking duration shown on Google Maps. Also, include links to the match details page and each hotel's details page."
544task-4e58a6,"I'm planning to visit the Getty Center in Los Angeles to check out their current special exhibitions.545 546First, please go to their official website and find a special exhibition with 'Future' or 'Science' related concepts in its name. If no such exhibition is available, find the most popular one. Confirm its specific start and end dates.547 548After obtaining these dates, please search for hotels near the Getty Center on Booking.com. The booking should be for an upcoming arbitrary weekend during this exhibition period (two nights, check-in Friday, check-out Sunday). I'm looking for three hotels with a rating of 8.5 or higher and marked with 'free cancellation'.549 550Finally, use Google Maps to determine the driving time from each of these three hotels to the Getty Center. Select one with a driving time of 15 minutes or less, and provide me with the hotel name, price, and driving time."
551task-4fe5fb,"I'll be in New York this upcoming weekend and would like to find an affordable music performance to attend (under $30).552 553Please search Eventbrite for events scheduled for this upcoming weekend. Select a band or singer that seems interesting.554 555To ensure the quality of the live act, after identifying a performer, search for their live performance videos uploaded within the last year on YouTube (exclude official music videos; focus on actual live footage).556 557Identify the live video with the highest view count. Then, provide me with the name of the performance, the Eventbrite link, and the title and view count of that specific YouTube video."
558task-502566,"I’m planning to move to Tampa, Florida in the near future and I’m very concerned about neighborhood safety. First, go to the Superfund page on EPA.gov and search for the **“Kassauf-Kimerling Battery Disposal”** contaminated site in Tampa. Obtain its full street address and current NPL status.  559 560Next, go to Census.gov QuickFacts and check the latest population estimate for **“Tampa city, Florida.”** Compare it with the value from the previous statistical period to confirm whether the population is trending upward.561 562After gathering that information, go to Redfin.com to find listings. Filter criteria: Tampa city, **House** property type, price **$450,000–$600,000**, at least **3 bedrooms**, and status must be **Active** (exclude Pending/Contingent). Try to find **3** listings that meet these criteria; if fewer than 3 are available, return all currently available matches and indicate the total count.  563 564To ensure safety, use Google Maps to calculate the **driving distance** from each listing to that battery disposal contamination site, and prioritize homes that are **more than 3 miles** away. If all are under 3 miles, still report the exact distances truthfully.565 566Output: the contaminated site address, NPL status, and Tampa’s latest population; plus for each listing: address, price, Redfin link, driving distance to the contaminated site, and whether it meets the “>3 miles” criterion."
567task-50f5d7,"Our project plans to introduce an NPM package named 'visx' for data visualization. However, given the recent surge in supply chain attacks, a security review is required before deployment.568 569First, search for the 'visx' package on NPM. Record its weekly downloads, latest version, number of dependencies, and whether it has Provenance verification. Then, navigate to its GitHub repository.570 571On GitHub, check the Commit activity over the past 3 months (number of commits), the number of Contributors, and whether there are any security issues labeled 'security' or 'CVE' in the Issue list (if any, record the Issue title and number).572 573Next, extract core algorithm keywords for the package (if specific visualization technology names are mentioned in the README or description). Search for related papers on Semantic Scholar, sort by citation count in descending order, and identify the top 3 most cited papers. Record the paper titles, citation counts, publication years, and whether a PDF is available.574 575Finally, use the core keywords to search Google Patents for related patents (specifically those with a 'Granted' status). If any are found, record the patent title, applicant, and grant date.576 577**Output:**578*   **NPM Package Information:** (Weekly Downloads, Latest Version, Number of Dependencies, Provenance Status, NPM detail page link)579*   **GitHub Health:** (Number of commits in the past 3 months, Number of Contributors, Presence of security issues and their IDs, GitHub repository link)580*   **Academic Endorsement:** (Titles of the 3 most cited papers, Citation counts, Publication years, PDF availability, Semantic Scholar link)581*   **Patent Risk:** (Existence of relevant granted patents; if yes, provide patent title, applicant, grant date, Google Patents link)582*   **Overall Risk Rating:** (Low/Medium/High, based on a comprehensive assessment of security issues, maintenance activity, academic credibility, and patent risk)."
583task-522faf,"I’m planning to drive an EV from Redding, California to Eugene, Oregon over a weekend next month. Since winter low temperatures can significantly reduce range, I need a dynamic charging plan.584 585First, use Google Maps to map this route, note the total mileage, and select 3 major intermediate cities along I-5 as potential stops (e.g., Mt. Shasta, Ashland, Roseburg).586 587Next, check the **Daily** weather forecast on AccuWeather for those 3 cities for the **departure day**. Focus on the **RealFeel® high temperature** and whether any precipitation/snow is forecast. Rule: if a city’s RealFeel® high is below **45°F**, or if the forecast includes any form of precipitation (**Rain/Snow/Ice**), mark that city as **“High Energy Consumption Risk.”**588 589For cities marked as **“High Energy Consumption Risk,”** return to Google Maps and find a nearby Tesla Supercharger with a rating above **4.5**. If no Tesla Supercharger above 4.5 exists in that city, choose the highest-rated Tesla Supercharger within the city and note **“Below 4.5”** in the output.590 591Output required:592- Total trip distance  593- For each of the 3 cities: city name, AccuWeather RealFeel® high on departure day, precipitation forecast details, risk status (Yes/No)  594- For each risk city: selected Supercharger name, rating, and Google Maps link"
595task-543362,"I am currently conducting cross-border product selection research for beauty products.596 597First, navigate to the official Sephora US website. Under the 'Skincare' category, identify products marked as 'Best Sellers'. Select 10 products with a rating of 4.3 or higher. For each selected product, record its product name, brand, Sephora price, and rating.598 599Next, use the format ""Brand + Product Name"" as keywords to search for the corresponding products one by one on Amazon. Record the Amazon price, rating, and number of reviews for each product.600 601Then, search for these same products on Target using the identical ""Brand + Product Name"" keywords. Record the Target price and stock status (in stock/out of stock).602 603Finally, visit the official brand website for each product (e.g., Drunk Elephant, La Mer, CeraVe). Locate the official product page. Record the official retail price from the brand's website and note whether there are any website-exclusive sets or special offers.604 605For each product, output the following information:606Product Name, Brand, Sephora Price, Sephora Rating, Sephora Product Link, Amazon Price, Amazon Rating, Amazon Number of Reviews, Amazon Product Link, Target Price, Target Stock Status, Target Product Link, Brand Official Website Price, Availability of Website-Exclusive Sets, Brand Official Website Product Link."
607task-550c46,"I'm a photography enthusiast planning to purchase a used Sony Alpha a6400 camera body by the end of 2025.608 609First, go to BestBuy.com and search for the Sony Alpha a6400 camera body. Use the filtering options to show only 'Open-Box Excellent' condition items, with a price under $800. Sort the results by 'Price: Low to High' and identify the cheapest one to establish a benchmark price.610 611Using this benchmark price, go to eBay.com for price comparison. Search for the same camera body ('Body Only'). Filter for 'Buy It Now' listings and 'Top Rated' sellers to ensure reliability. Find three options that are cheaper than the BestBuy benchmark price.612 613Finally, considering that used batteries often don't last long, go to Amazon.com and search for a third-party 'dual battery charger kit' compatible with the a6400 (NP-FW50 model). The kit must have a rating of 4 stars or higher. Sort the results by 'Avg. Customer Review' and select the one with the highest rating and over 1000 reviews.614 615Output: The BestBuy benchmark price; the prices and seller names for the three cheaper eBay options; the full title, price, and number of reviews for the selected Amazon kit; and the direct product page links for the chosen items from all three websites."
616task-55d8ab,"I plan to assemble a 4K 144Hz gaming PC with a budget of $2500. The core hardware list includes an RTX 4080 graphics card, i9-14900K CPU, Z790 motherboard, 32GB DDR5 RAM, and a 2TB NVMe SSD.617 618First, search for these five hardware components separately on Best Buy, Amazon, and eBay. Record the price and stock status of each component on all three platforms, then filter for available options.619 620Next, go to the Reddit Gaming community and search for keywords related to this build (e.g., 'RTX 4080 i9-14900K build'). Find 2-3 recent discussion threads to evaluate opinions on the compatibility and price-performance ratio of this configuration, paying special attention to any recommendations to avoid potential issues (pitfalls).621 622Finally, find a suitable desk at IKEA for the PC setup. Requirements: width at least 120cm, weight capacity at least 50kg, and price not exceeding $300.623 624Output:625For each hardware component: Name, Best Buy price and stock, Amazon price and stock, eBay price and stock, and product page links for each platform.626For Reddit posts: Title, link, and a summary of core evaluations (compatibility/price-performance/pitfall advice).627For the IKEA desk: Product name, dimensions, weight capacity, price, and product page link."
628task-5648f1,"**Task Description:**629 630I want to upgrade my live streaming audio equipment, aiming for the sound quality level of streamer **shroud**.631 632First, navigate to his Twitch channel homepage. By consulting the 'Panels' or 'About' section (which might be located below the chat or in a dedicated tab), identify the specific model of microphone he currently uses.633 634Once the model is identified, go to Amazon and search for this microphone, specifically looking for a **new**, in-stock listing with a rating of 4.5 stars or higher. Record its price.635 636Next, to save money, search for the same microphone on eBay. Use the filtering options to display only 'Used' or 'Refurbished' condition items, and ensure the seller is a 'Top Rated Seller' to ensure a secure transaction. Identify the 3 lowest-priced options (including shipping) on eBay that meet the above criteria.637 638Finally, output: The complete microphone model, the new price on Amazon, the Amazon product link, the respective prices, condition, seller rating, and corresponding links for the 3 eBay options, and calculate how much money would be saved by purchasing the lowest-priced eBay option compared to the new price on Amazon."
639task-5686a2,"I am considering purchasing a used Toyota Sienna (2021-2024 models) and would like to identify any common issues or recurring problems specific to these model years.640 641First, search for Sienna reviews on Car and Driver. Focus on finding a 'Long-Term Road Test' or a comprehensive review published after 2021. Identify and record at least two significant drawbacks ('Lows') mentioned by the editors.642 643Next, search Reddit for real-world feedback from owners concerning these two identified drawbacks. Locate the most discussed (highest comment count) relevant posts and summarize the solutions or workarounds proposed by the community.644 645Finally, to potentially verify these issues firsthand, use Google Maps to find the official Toyota dealership closest to downtown San Francisco, CA, and calculate the estimated driving time to reach it.646 647Output: The drawbacks mentioned by Car and Driver along with the article link; For the highest-engagement Reddit post, provide its title, comment count, link, and a summary of the main solutions discussed; The dealership's name, address, distance from downtown San Francisco, and the estimated driving time."
648task-58a3d0,"I'm a long-time follower of Tesla and recently learned that their autonomous driving technology (FSD/Autopilot) is under considerable regulatory scrutiny.649 650First, please navigate to the official SEC website to find Tesla's latest annual report (10-K). In the 'Risk Factors' section, carefully identify and extract the key sentences where the company officially describes risks related to 'regulatory or legal proceedings concerning autonomous driving technology'.651 652Next, take this official statement and search on Quora for a highly engaged, relevant question (e.g., with numerous answers or followers) to see if the general public or industry professionals are discussing similar concerns.653 654Finally, please compare: are there any significant discrepancies between the official phrasing in the financial report and the central topics of discussion among Quora users?"
655task-59ce5f,"I need to select a sentiment analysis model for my company's FinTech project.656 657Please go to Hugging Face and search for the keyword 'financial sentiment'. Use the left-hand filter bar to set the task to 'Text Classification' and library support to 'PyTorch'. Sort by 'Most Downloads' and select the top 3 models with the highest download counts.658 659For each of these 3 models, navigate to its detail page to find the associated GitHub repository link. After navigating to GitHub, click to view the commit history, and record the specific date of the most recent commit on the main branch to assess maintenance activity. Additionally, look for the arXiv paper link referenced by the model, either on the HF model card or in the GitHub README.660 661Output the following information for these 3 models: Model Name, HF Detail Page Link, HF Download Count, GitHub Repository Link, GitHub Most Recent Commit Date, arXiv Paper Link (if none, state 'N/A')."
662task-5a4d4e,"I recently want to arrange music in the style of Bill Evans, and I need to find some specific references and advanced tutorials.663 664First, please go to Spotify and search for the official playlist 'This Is Bill Evans'. Tell me the current number of 'Saves' for this playlist and the title of the first track on the list. I need to confirm the style.665 666After confirmation, please go to Udemy to find courses for me. Search for 'Jazz Piano Improvisation'. To ensure quality, please filter the courses in the sidebar by a rating of 4.5 or higher, and select 'Intermediate' or 'Expert' for the difficulty level. I need to confirm whether the course outline teaches the specific technique of 'Rootless Voicings'.667 668Please find a course with either a 'Bestseller' or 'Highest Rated' label. Expand its 'Course content' section and carefully examine the titles of the lectures/lessons. Once you find a course that includes this technique, send me the course link and excerpt the name of the chapter or lesson where this technique is mentioned."
669task-5d022c,"I am developing a pixel-art action Roguelike game and would like to research the map generation mechanisms of benchmark products.670 671First, go to Steam. Use the filtering function to find games tagged with both 'Pixel Graphics' and 'Action Roguelike'. To ensure quality, sort the results by 'User Reviews', and identify the top-ranked game that is paid (not free). Go to its detail page and confirm if its positive review percentage exceeds 95%.672 673Once confirmed, I want to find similar open-source implementations. Go to GitHub and search for 'Roguelike dungeon generation', filter the language to C#, sort by 'Stars', and identify the top-ranked repository.674 675Finally, compile for me the name and positive review percentage of the Steam game, as well as the URL and star count of the GitHub repository."
676task-5df385,"Our team is preparing to transition from Vue to React and would like to purchase a few physical books for the office library corner. Please help me find three React technical books on Goodreads, published after 2022, with a rating of 4.3 or higher. The content must be in-depth; avoid purely introductory books. Once you've identified suitable books, search for them on Google Books. If a preview is available, check the table of contents or main text to determine if the code examples are written using functional components + Hooks, or the older Class components. Finally, please provide me with the titles, authors, exact Goodreads ratings, and confirmation of Hooks-based examples for these three qualifying books."
677task-5e833f,"I'm interested in researching the phenomenon where film critics' opinions and audience preferences diverge significantly.678 679First, navigate to the 'Movies' section on Rotten Tomatoes. Find a film currently 'In Theaters' or available 'At Home' (streaming) that meets the following criteria: a high Tomatometer score from professional critics (over 80%) but a low Audience Score (under 50%).680 681Once you find a suitable movie, please tell me its name and the exact Tomatometer and Audience Scores.682 683Next, search for this film on IMDb. Go to its 'User Reviews' section, filter for reviews rated 1 or 2 stars, and sort them by 'Helpfulness'. Summarize the top-ranked lengthy review, identifying the primary reasons for audience dissatisfaction."
684task-5ea2c0,"I am a full-stack developer, and I want to learn a new AI tool towards the end of the year. Please go to the GitHub Trending page and filter for the most popular Python projects for 'this month'. Carefully examine the top-ranked projects. Identify the first project whose description explicitly contains either 'Agent' or 'LLM' as keywords, and which also has a `requirements.txt` file present in its repository's root directory. Once found, search on YouTube for practical tutorials for this project. Please ensure to use the filter function to only show videos released 'this year' and with a duration exceeding 4 minutes (to ensure they are not short trailers). Identify the one with the highest view count. Output: GitHub project name, GitHub project link, Star count, direct link to the `requirements.txt` file, YouTube video title, video duration, video publication date, video view count, YouTube video link."
685task-5fe6bf,"I am a product manager at an AI startup, and my team is evaluating the feasibility of commercializing **Monocular Depth Estimation** technology as an API service. Please help me conduct a technical commercialization research task:686 687First, go to Semantic Scholar and search for **""monocular depth estimation""**. Sort by citation count and identify the 3 most-cited papers published since 2024. Record each paper’s title, authors, publication year, citation count, and TLDR summary.688 689Then open each paper’s detail page and check whether it includes a GitHub link. If it does, visit the corresponding GitHub repository and review the README to extract the GPU model and VRAM requirement needed for model inference (e.g., RTX 3090 / 24GB). If the README does not explicitly state GPU requirements, check the repository’s requirements, docs, or issues for minimum hardware information. If still not found, mark it as **“GPU requirements not explicitly specified.”**690 691Next, search Amazon for those GPU models, filter for **new** condition, and prioritize products sold by **Amazon**. Find the currently lowest-priced option and record the GPU model, price, seller type, and product link. If no Amazon-sold listing is available under new condition for that model, select the lowest-priced non-Amazon seller listing under new condition as a fallback, and clearly label it as a fallback choice.692 693Finally, go to Google Scholar and search for **""depth estimation API pricing""** or **""3D reconstruction API cost""** to find pricing clues for existing competitor services (e.g., Replicate, Hugging Face Inference API). Then visit each service’s public pricing page to verify pricing details. If found, record the provider name, pricing model (per call / monthly subscription), and exact price.694 695**Output required:**696- For each paper: title, authors, publication year, citation count, TLDR summary, whether a GitHub link exists, and Semantic Scholar paper link  697- Extracted GPU requirements from the corresponding GitHub repo (model and VRAM; explicitly note if unspecified), plus GitHub link  698- Lowest Amazon price for that GPU, seller type, and product link (and whether it is Amazon-sold)  699- Competitor service name, pricing model, and exact price (if not found, explain why and list attempted paths)"
700task-642cb8,"I’ve recently started learning machine learning and want to find a high-quality hands-on project tutorial on Bilibili to follow. Please search using keywords like “machine learning project practice” or “deep learning project,” sort by view count, and identify videos with at least 500,000 views.701 702For each candidate video, check whether the description includes a GitHub repository link. If the description is unclear, expand it to view the full text, and also check the pinned comment in case the uploader added the code link there.703 704After finding a GitHub link, verify whether the repository actually exists and whether the code appears complete. Check for files such as a README and requirements.txt, and see whether the repository has been updated recently.705 706If a video does not provide a GitHub link or the link is invalid, move on to the next video that meets the view-count requirement and continue until you find at least one accessible repository. If none can be found in the end, clearly state the scope of what was checked and why no valid result was found.707 708Finally, tell me whether the video is worth following for learning and whether the code can be used directly."
709task-6519b0,"I am evaluating the investment potential for vacation rentals in Kissimmee, Florida (34747), with a baseline for analysis set for December 2025. First, on Zillow, find three single-family homes for sale. Requirements: price between $450,000 and $600,000, at least 4 bedrooms, and explicitly listed with a 'Private Pool' in the 'Facts and features' section.710 711For each of these three Zillow properties, find comparable Airbnb listings to estimate potential rental income. Using the map search function, within the same neighborhood or a 0.5-mile radius of each Zillow property, find a short-term rental property that also has 4 bedrooms and is tagged as 'Guest favorite'. Check the nightly rate and cleaning fee for this comparable listing for a 5-night period, specifically from March 15th to 20th of the next calendar year (e.g., if the current date is in 2024, use March 15-20, 2025).712 713Finally, using Google Maps, calculate the driving time from each Zillow property to 'Magic Kingdom Park' and verify if it is within a 20-minute drive.714 715**Output: Zillow property address, sale price, Zillow link, Airbnb comparable listing link, Airbnb nightly rate (for the specified March 15-20 period), cleaning fee, whether it is a 'Guest Favorite', and driving time to Magic Kingdom Park.**"
716task-662113,"I am a beauty blogger and I want to purchase a batch of foundations for a review video. I need to find products with high cost-effectiveness.717 718First, go to Sephora and search for 'foundation'. Sort the results by 'Best Selling' and identify the top 5 foundations from different brands. Record the brand, full product name, Sephora price, and rating for each.719 720Then, take the full names of these 5 products (brand + product name) and search for them on Amazon and Target, respectively. Extract the price and inventory status from each platform. For Amazon, note whether there is a 'Prime' badge and an 'Official Store' indicator. For Target, check if there are any 'Target Circle' member discounts.721 722Finally, visit the official websites of these 5 brands one by one (e.g., Fenty Beauty, NARS). On each official website, search for the corresponding product or navigate to the 'foundation' category. Find the official website price for the product, check for any current promotional activities (e.g., 'Buy 2 Get 1 Free', 'Gift with Purchase'), and look for links to bundle deals.723 724For each foundation, output the following:725*   Brand726*   Full Product Name727*   Sephora Price and Rating728*   Sephora Product Page Link729*   Amazon Price and Prime Eligibility (yes/no)730*   Amazon Product Page Link731*   Target Price and Circle Offer Information732*   Target Product Page Link733*   Brand Official Website Price734*   Description of Current Promotions on the Official Website735*   Official Website Product Page Link"
736task-666eae,"I'm working on a Named Entity Recognition (NER) project and found the `dslim/bert-base-NER` model on HuggingFace to be effective. I need help understanding the research team behind this model.737 738Please perform the following steps:7391.  Go to HuggingFace, locate the `dslim/bert-base-NER` model, and extract the associated paper link and author information from its Model Card.7402.  Navigate to arXiv, find the paper, and note down the full name of the first author and their institutional affiliation.7413.  Search for this first author on Semantic Scholar. Identify their top 5 most cited papers, listed in descending order of citation count.7424.  Finally, search GitHub for the author's username (you can use their name and affiliation as keywords). Find their personal profile and list at least their top 3 open-source projects by star count.743 744Output format:745*   Model Name746*   HuggingFace Model Link747*   Paper Title748*   arXiv Link749*   First Author's Full Name750*   Author's Affiliation751*   Author's Semantic Scholar Profile Link752*   Author's 5 Most Cited Papers (for each: Paper Title, Publication Year, Citation Count, Paper Link)753*   Author's GitHub Username754*   GitHub Profile Link755*   Author's 3 Highest-Starred Projects (for each: Project Name, Star Count, Project Link)"
756task-683779,"I’m a fan of Frank Herbert’s original *Dune* novel and want to know how much the newly released *Dune: Part Two* has changed it. Please do an in-depth comparison for me.757 758First, go to Goodreads and find Frank Herbert’s original *Dune*. I want to see the latest discussion, so sort the reviews/comments by **“Newest.”** Find one recent, detailed long review that explicitly mentions “movie” or “film,” and note the user’s core viewpoint.759 760Then go to Letterboxd and find *Dune: Part Two*. Browse several pages of reviews, and/or expand collapsed spoiler content. Prioritize finding **two reviews** that explicitly compare the film with the **book**. If there are fewer than two explicit book-to-film comparison reviews, you may supplement up to two total by selecting in-depth reviews that clearly discuss **adaptation trade-offs / faithfulness / omitted or changed plot points**. In your results, label which ones are direct book-film comparisons and which ones are adaptation-focused discussions.761 762Finally, summarize the main points of disagreement between hardcore audiences on these two platforms regarding this adaptation, and list the user IDs of the reviewers you cited."
763task-6909ef,"I plan to relocate with my family to North Carolina within the next year, and I’m currently deciding between downtown Raleigh (ZIP 27601) and downtown Charlotte (ZIP 28202). Environmental health is my top priority. First, go to epa.gov and use the ECHO tool to find, for each of these ZIP code areas, the number of facilities that have had a **Significant Violation** in the past three years. Then use the **How’s My Waterway** tool to check the overall **Condition** of the major water bodies in each area.764 765Compare these two sets of results and select the ZIP code area with fewer significant-violation facilities and better water quality conditions. Then go to Zillow and find 3 for-sale **Single Family Homes** in the selected ZIP code area, with at least 3 bedrooms, priced at no more than **$650,000**, sorted by **Newest**.766 767If fewer than 3 listings are available with the above filters, keep property type = Single Family, bedrooms ≥ 3, and sorting = Newest unchanged, and first relax the price cap (to **$700,000**). If there are still fewer than 3, return all available listings in that ZIP code under the current criteria and clearly state the final number found.768 769**Output requirements:** Provide a comparison of significant-violation facility counts for the two ZIP codes, plus the names and conditions of the major water bodies. For listings in the selected area, include address, price, bedroom count, days on Zillow (or listing/publish date), and Zillow detail-page link, and indicate whether each listing meets the “≤ $650,000” requirement."
770task-6ab559,"I am currently preparing for the 'Fast-Paced Reading Marathon' event scheduled for January 2026, and I need your assistance in compiling a booklist.771 772Please begin by visiting TheStoryGraph's 'Browse' or recommendation page. Utilize the filters to locate 3 novels that satisfy all the following conditions:7731.  **Moods**: Must simultaneously include both 'Adventurous' and 'Funny'.7742.  **Pace**: Must be 'Fast-paced'.7753.  **Page Length**: Use the page length slider to restrict the range to under 350 pages.776 777Identify the three highest-rated books that meet these criteria.778 779Next, search for each of these three books individually on OpenLibrary.org. Verify if a digital version is available for direct borrowing (indicated by a 'Borrow' button) or online reading (indicated by a 'Read' button).780 781Please output the following information for each book: Book Title, Author, StoryGraph Rating, StoryGraph Page Count, specific availability status on OpenLibrary (e.g., Borrow/Read/Not Found/Checked Out), and the direct links to the book's detail pages on both platforms."
782task-6de61d,"I want to visit that famous cyberpunk-style residential building in Chongqing, but I don’t know its exact name. First, go to Bilibili and search for “Chongqing cyberpunk residential building” (重庆 赛博朋克 居民楼). Find a video with relatively high views, then check the comments or video description to identify the building’s exact name (it should be three Chinese characters).783 784After obtaining the name, go to Airbnb and locate bookable listings around that area. I plan to stay next weekend (Saturday to Sunday), for 2 people, with a budget of no more than RMB 400 per night. Be sure to use map view to find 3 listings within walking distance of that building, preferably with river views.785 786Provide the names, prices, and ratings of these 3 listings.787 788If there are no bookable Airbnb listings in that area, or if there are access/listing restrictions, record the reason for the limitation, then search on Airbnb in a nearby alternative district (e.g., Jiefangbei) and complete the same filtering process."
789task-6e79f1,"We are preparing to hire a Senior Data Analyst in San Francisco and need to develop a competitive compensation package.790 791First, go to LinkedIn and search for current job openings for ""Senior Data Analyst"" in the San Francisco Bay Area. Browse through several pages and select 3 technology companies that clearly state a salary range. Record their company names, salary ranges, and check if the job descriptions mention any special benefits (e.g., stock options or remote work).792 793Next, for these 3 companies, go to Glassdoor and review genuine employee feedback. Remember to filter the reviews to only show ""Current Employee"" comments, and specifically summarize their key complaints or criticisms regarding ""Management.""794 795Finally, go to BLS.gov and find the official data for the broader category ""Data Scientists"" (which typically includes analysts). I need to know the ""Projected Growth"" for the next ten years (2023-33) to determine if we are hiring for a growing profession."
796task-6f5fc0,"I’m writing a report on global dengue transmission trends and need your help with some research. First, go to Google Scholar and search for epidemiological studies on dengue from the past year. Find several papers with relatively high citation counts, prioritizing information that can be obtained directly from titles and abstracts, and identify the high-risk regions and transmission patterns mentioned in those papers. If a result requires institutional login or payment to access the full text, note that access restriction and move on to another paper with publicly available abstract information.797 798Then, go to Our World in Data and check global dengue data to see which countries or regions have had the fastest growth in case numbers in recent years, and whether that aligns with the high-risk areas identified in the papers. If a map view is available, that would be even better for visualizing geographic distribution.799 800Finally, help me summarize whether the conclusions from academic research are consistent with actual data trends, and highlight any new findings or contradictions."
801task-6fd2dc,"I recently watched ""Interstellar"" and was deeply moved by Hans Zimmer's soundtrack, especially the organ piece. I'd like to find the complete soundtrack album for the movie to listen to, and also explore other classic works by Hans Zimmer. Additionally, I recall that there are musicians on Bilibili (B站) who have specifically analyzed the soundtrack composition techniques for this film. Please help me find analytical videos with high view counts and in-depth discussions, preferably content that is over 10 minutes long and offers deep insights, as I want to systematically learn how film scores enhance emotional atmosphere."
802task-72b9f2,"I work in AI video generation and recently want to catch up on new technologies in the Text-to-Video domain.803Please help me search for models in this field on Hugging Face, filter them by the PyTorch framework, and sort them by their most recent update time.804Among the first two pages of results, identify the model with the highest download count. (Important: exclude models that are not solely Text-to-Video but also include other complex task tags; focus only on models dedicated to Text-to-Video).805Once found, navigate to its Model Card and check if it references an ArXiv paper. If it does, proceed to ArXiv to locate that paper. Pay particular attention to the core architecture mentioned in the Abstract (e.g., Diffusion-based or Transformer-based).806Finally, report the model name, paper title, and core architecture."
807task-72d45e,"I'm looking for an engaging RPG on Steam to pass the time. Please go to the Steam store and search for ""RPG"". Use the filters on the left to only show games priced under 50 RMB and with ""Overwhelmingly Positive"" reviews. Browse through several pages and select at least 5 promising indie games that are not DLCs. Then, for each selected game, go to the HowLongToBeat website and look up their ""Main Story"" completion time. Finally, filter out games with a ""Main Story"" completion time exceeding 20 hours. List the game title, its current price on Steam, and its ""Main Story"" completion time from HLTB."
808task-73468a,"I’m planning to produce an in-depth analysis video on an indie puzzle game, around 20 minutes long. First, help me find a game on Steam tagged **indie + puzzle** that was released within the past 12 months, has an **Overwhelmingly Positive** rating, and has **more than 1,000 reviews**.809 810Then, check this game on Metacritic and record its **Metascore** and **User Score**, including both the numeric scores and their rating labels.811 812Next, search Fandom for the game’s Wiki and go to a **Design** or **Gameplay**-related section to review detailed explanations of its mechanics. If there is no Fandom Wiki for this game, or no relevant Design/Gameplay section, use the game’s **Steam Community Guides** page or an **official/developer-published gameplay explanation page** as an alternative source, extract the key mechanics points, and clearly note the replacement source link.813 814After that, go to YouTube and search for **[game name + gameplay]**, filter for videos longer than 10 minutes, sort by view count, and pick the top 2 high-view gameplay videos. Then search **[game name + analysis]** or **[game name + review]**, again filter for videos longer than 10 minutes and sort by view count, and select the top 2 in-depth commentary videos.815 816Finally, go to Bilibili and search for the game’s Chinese or English name plus **解说 / 评测** (commentary/review), filter for videos longer than 10 minutes, sort by play count, and pick the top 2 Chinese commentary videos.817 818Output the following:819- Game title  820- Steam review status (e.g., Overwhelmingly Positive)  821- Steam review count  822- Steam link  823- Metacritic Metascore (numeric)  824- Metascore rating label  825- Metacritic User Score (numeric)  826- Metacritic link  827- Summary of the game’s core mechanics/design highlights (preferably from Fandom; if an alternative source is used, explicitly label it)  828- Mechanics information source link  829- Titles and links of 2 YouTube gameplay videos  830- Titles and links of 2 YouTube analysis/review videos  831- Titles and links of 2 Bilibili commentary videos"
832task-73c4b7,"Our team plans to implement a speech cloning system in early next year and currently needs to conduct a technology selection. Please conduct this research as of now.833 834First, search for ""Text-to-Speech"" models on HuggingFace. Using the left-hand filter bar, only consider models that support the ""PyTorch"" framework and have an ""Apache-2.0"" License (to ensure commercial viability). Sort by download count and select the top 3.835 836Next, find the corresponding GitHub repositories for these 3 models. Navigate to their Issues page, count the number of Open and Closed Issues for each, and calculate the Open/Closed ratio.837 838Finally, for the model with the lowest Open/Closed ratio (i.e., highest resolution efficiency), search on Bilibili for currently available, detailed Chinese tutorial videos longer than 10 minutes (with keywords including ""fine-tuning"" or ""deployment""). Find 3 videos that meet these criteria.839 840Output: The names of the top 3 models, their HuggingFace download count, HuggingFace update date, GitHub link, and GitHub Open/Closed issue counts; and for the finally selected model, the titles, durations, publication dates, and links of the Bilibili tutorial videos."
841task-74524f,"As a plant photography enthusiast, I plan to visit Kew Gardens in London to photograph rare tropical Orchidaceae.842 843First, on Kew Gardens' official scientific database POWO (powo.science.kew.org), search for plants with the genus ""Bulbophyllum"". From the results, identify 3 species that are ""Native to"" ""Madagascar"" and have ""Images"". For these 3 plants, query Wikipedia to find their ""Common name"" (if none, mark as such) and whether they are ""Epiphyte"".844 845Next, using Google Maps and starting from Kew Gardens, plan a walking tour route: first visit the ""Princess of Wales Conservatory"" (where such plants are typically displayed), then proceed to the nearby ""The Hive"" art installation. Finally, find a cafe within a 10-minute walk from ""Kew Gardens Victoria Gate"" with a rating above 4.5 for a break.846 847Output the scientific name, native origin, Wikipedia link, and epiphyte status for each of the 3 plants, along with the total distance of the walking route and the name of the cafe."
848task-745653,"I am currently redecorating my living room and would like to draw inspiration from IKEA designs, but I have a limited budget.849 850Please navigate to the 'Ideas' or 'Inspiration' section of the IKEA website and find a living room showcase or room set with a 'Modern' or 'Minimalist' theme. From this showcase, identify the three most prominent/core large furniture items (e.g., a sofa, coffee table, and rug). Record their names, IKEA prices, and main dimensions.851 852Next, search on Target and Amazon, respectively, for alternative/substitute versions of these three furniture items. The requirements for these substitutes are: their appearance and style should be similar (e.g., both are grey fabric sofas), their dimensions must be within a 10cm tolerance, and their price must be at least 10% cheaper than the original IKEA version.853 854Finally, compile a detailed list presenting these three comparison sets (IKEA item vs. substitute item), and calculate the total savings if all substitute items are purchased."
855task-747160,"I’m looking for safe facial moisturizers for sensitive skin during pregnancy and want to verify that the ingredients are truly safe and reliable.856 857First, go to the EWG website. Under the **Face & Body** section, find the **Facial Moisturizer** category, then filter for products with an **EWG score of 1–2** (green safety range) and an **EWG VERIFIED** badge. Try to find and record up to **5 products**.858 859Next, search for these same products on **Sephora** and collect, for each one: user rating, price, whether it has the **Clean at Sephora** label, and especially whether reviews mention feedback from users with sensitive skin or who are pregnant.860 861Then search for the same products on **Target**, compare prices and **Target Circle** member offers, and determine which platform is more cost-effective.862 863Finally, visit each product’s official brand website and locate either the certification page or product detail page to confirm whether the site displays certifications such as **EWG VERIFIED**, **USDA Organic**, or **Cruelty-Free**.864 865For each product, output:866- Full product name  867- EWG score (specific number: 1 or 2)  868- EWG VERIFIED certification (Yes/No)  869- Sephora rating and price  870- Clean at Sephora label (Yes/No)  871- Target price and Target Circle offer details  872- Certification labels shown on the brand’s official website  873- Detail page links from each platform (**one each from EWG / Sephora / Target / brand official site**)874 875If fewer than 5 products meet the EWG criteria, proceed with the actual number found. If a product cannot be found on Sephora or Target, note that and continue collecting the remaining information for other products."
876task-76041b,"I am planning to move to the South Lake Union area in Seattle, as I've heard it's suitable for young professionals. First, please help me search for apartments in this area on Apartments.com. My budget is within $3500 per month. I need a 2-bedroom unit with an in-unit washer & dryer. Please filter out the top 3 apartments that meet these criteria and note down their specific addresses.877 878Next, I'd like to check the convenience of living. For each of these three addresses, please go to Yelp and see if there are any coffee shops with a rating of 4.0 or higher within walking distance (search within 0.5 miles).879 880Finally, please provide a comparative summary: List the names of these three apartments, their monthly rent, and for each, the name and Yelp rating of the highest-rated coffee shop nearby. If there are no eligible coffee shops around a particular apartment, please state that as well."
881task-762d43,"I'm a student in NYC planning an upcoming weekend trip to see the band 'The National'. First, check Spotify's concert listings to find a show scheduled for an upcoming Friday or Saturday. Select a city that seems close to New York (e.g., Philadelphia, Boston, DC). Once a target city is found, go to Google Maps and calculate the driving directions from 'New York, NY' to that city's downtown to confirm the driving distance is under 250 miles. Next, look for accommodation on Expedia: search for hotels in that city for the night of the concert (i.e., the evening of the selected Friday or Saturday). Use the interactive map view to pick a hotel that is rated 4.0 or higher and costs under $250. Finally, check Eventbrite for other 'Music' events in that city for the following day (i.e., Saturday or Sunday), specifically looking for an event marked as 'Selling fast' or 'Promoted' to round out the weekend. Output: The concert city and date, Google Maps driving distance, the chosen hotel's name, price, and rating, and the Eventbrite event name with its status tag. Provide URLs for the Spotify artist page, Google Maps route, Expedia hotel detail, and Eventbrite event."
882task-786abc,"I have an onsite interview next week at a tech company in Seattle and need to book a 3-night stay.883The company is Amazon Web Services. First, please confirm their specific headquarters address in Seattle on LinkedIn.884Then, search for Airbnb listings near that address. The check-in dates should be next Monday to Thursday. Find 3 options that have high ratings and a convenient commute to the company.885I'd also like to understand the amenities around these listings, as I might move there if I get an offer. Therefore, I want to know if the area is suitable for long-term living.886Finally, please provide a comparison of these 3 listings, including: distance from the company, commute time, nightly price, rating, and nearby amenities like restaurants and supermarkets."
887task-79f5e0,"I want to seriously learn watercolor painting on Udemy. Please find me one beginner course with these core requirements: a rating of at least 4.5 and more than 1,000 reviews. Also, prioritize courses with the “Bestseller” badge. If no course meets both the core requirements and has a Bestseller badge, then choose from courses that meet only the core requirements.888 889After selecting the course, carefully review its course description or “Requirements” section. The instructor may specify exact paper brands (e.g., Arches) or specific brush models. If no exact brand/model is provided, extract actionable general specifications instead (e.g., 100% Cotton Paper, 300gsm, Cold Press, Round Brush sizes).890 891Then extract the key painting materials required by the instructor, search for the same items or corresponding specs on Amazon, and filter for products rated above 4 stars.892 893Finally, provide me with a checklist that includes:894- Course link  895- Instructor-required material names  896- Corresponding Amazon product links, ratings, and prices  897- A note for each material indicating whether it is an “exact brand/model match” or a “matched by general specifications”898 899If any material has no Amazon result above 4 stars, keep that search item and mark it as: “No products found that meet the rating requirement.”"
900task-7a01b5,"I am a Senior Software Engineer in San Francisco, and in the current economic climate I care a lot about job stability. Please help me identify 3 job postings from publicly listed companies. First, search for “Senior Software Engineer” on Glassdoor, set the location to “San Francisco,” and use filters to show only roles with salaries above $170k. Try to select 3 positions that you can confirm are from publicly traded companies (if fewer than 3 are available, record the actual number found and continue with the remaining analysis). For each company, conduct the following risk review:901 9021. Go to the SEC website (sec.gov), find the company’s latest quarterly report (10-Q), and locate the “Item 1A. Risk Factors” section. Check whether there are risk disclosures related to “restructuring,” “workforce reduction,” or “cost efficiency,” and extract relevant summaries.9032. Return to Glassdoor and open the company’s Reviews page. Filter for “Current Employee” reviews and sort by “Lowest Rating.” Prioritize negative reviews from the most recent 6 months to identify what employees are mainly complaining about (e.g., management chaos, budget cuts, etc.); if there are not enough samples in the past 6 months, expand to the most recent 1 year and note that adjustment.904 905If you encounter a login wall or access restrictions during Glassdoor steps, record the blocked steps and reasons; for companies that remain accessible, continue completing the SEC review and other executable steps.906 907Output requirements: company name, job title, salary range, job link, SEC 10-Q filing link, SEC risk-factor summary, Glassdoor negative-review page link (if accessible), and summary of current-employee negative reviews (if accessible)."
908task-7ae3f0,"I've recently received an offer from Tableau in Seattle, with the office situated on North 34th Street. I plan to start in early 2026 and need to find accommodation.909 910First, please confirm the exact address of Tableau Software's headquarters in Fremont, Seattle.911 912Using this address as the destination, please plan a driving route on Google Maps for an 8:30 AM arrival on a Monday. Identify three specific neighborhoods where the peak-hour driving commute time falls between 25 and 35 minutes (inclusive) to be considered as prospective residential areas.913 914Then, on Apartments.com, for each of these three identified neighborhoods, locate one 1-bedroom apartment that has a rating of 4.0 or higher, a rent between $2500 and $3200, and explicitly features ""In Unit Washer & Dryer"" amenities.915 916Output should include: Tableau's exact address; the names of the three prospective neighborhoods along with their corresponding estimated Google Maps commute times; and for each recommended apartment: its name, address, rent, rating, and the Apartments.com detail page link."
917task-7af2a2,"I’m a die-hard Coldplay fan based in New York (JFK), and I want to travel to see their upcoming tour dates in Europe.918 919First, check Spotify for Coldplay’s tour schedule in Europe and select 3 different cities. If there are no available European dates on the current page, then find the 3 nearest upcoming tour cities worldwide (not limited to Europe) and clearly note this in the results.920 921Next, for each of those 3 cities, use Google Flights to search round-trip flights that align with the concert timing (arrive 1 day before, depart 1 day after), prioritizing nonstop options, and identify the city with the lowest airfare. If a city has no nonstop flights, you may choose the option with the shortest layover and note that explicitly.922 923Finally, for the winning city, use Airbnb to find accommodation for the concert period. Requirements: it must be labeled **“Guest Favorite”**, priced at **$300/night or less**, and verified in **map view** to be very close to the concert venue (walkable or a short drive).924 925Output: a flight price comparison across the 3 cities, plus the final selected Airbnb listing name, nightly price, estimated total trip cost (**flight + 2 nights of lodging**), and listing URL."
926task-7b62fb,"I am a researcher in digital preservation of cultural heritage, preparing a proposal report for the museum director on AI-based artwork restoration technologies. First, search Semantic Scholar for **“AI art restoration”** or **“neural network painting restoration”**, filter for papers published since 2024, sort by citation count in descending order, and identify the 5 most-cited papers that provide a PDF download link. Record each paper’s title, authors, publication year, citation count, and paper link.927 928Then, extract the core algorithm names from the abstracts or full texts of these 5 papers (e.g., GAN-based restoration, Transformer-based inpainting), and search GitHub for corresponding open-source implementations. Prefer projects with more than 500 stars and updates within the last year. Select the 3 most active projects and record the project name, star count, most recent update date, primary programming language, and project link. If fewer than 3 projects meet all criteria, keep the “updated within the last year” requirement, prioritize higher-star relevant projects, and clearly note which ones do not reach 500 stars. If GitHub access is restricted or requires login, document the limitation and continue by using code links from paper pages or searching via a search engine with queries like “algorithm name + GitHub” to find publicly accessible repositories.929 930Next, search Google Patents for patents related to these algorithms. Filter for granted U.S. patents filed after 2020, identify 2–3 patents, and record the patent number, title, assignee, filing date, and patent link.931 932Finally, visit the Getty Museum official website. Under **“What’s On”** in Current or Future exhibitions, check whether there are exhibitions or events related to digital restoration or technology-driven conservation. If found, record the exhibition name, date range, summary, and link. If none are relevant, search the Getty collections pages using the keywords **“digital”** or **“conservation”** to find cases of digitally supported restoration, and record 1–2 cases with object name, description, and link.933 934**Output format:**  935- 5 papers (title, authors, publication year, citation count, Semantic Scholar link)  936- 3 GitHub projects (project name, star count, latest update date, primary programming language, GitHub link)  937- 2–3 patents (patent number, title, assignee, filing date, Google Patents link)  938- Getty-related exhibition(s) or collection case(s) (name, description, link)"
939task-7c3fca,"I want to buy three types of upgrades for my 2025 Ford Bronco: a roof rack, a winch, and all-terrain tires. First, go to Car and Driver and find the review article for the 2025 Ford Bronco to see which upgrade brands and specific models they recommend. If that review does not provide specific brands/models for all three upgrade categories, continue searching within Car and Driver for Bronco upgrade guides or related articles to fill in the missing recommendations.940 941Then take those recommended models and search for them on Amazon, eBay, Target, and Best Buy to compare prices. Apply these filters: rating above 4 stars, in stock, and standard shipping or faster delivery available. For each upgrade category, identify the lowest-priced option across the four platforms; however, if the lowest-priced option has a rating below 4.5 stars or fewer than 50 reviews, choose the second-lowest-priced option instead.942 943Finally, output for each upgrade category: recommended brand and model, the most cost-effective platform to buy from, that platform’s price, rating, number of reviews, stock status, shipping method, and product detail page link."
944task-7d27af,"This Saturday I’ll be in Chicago and want to plan a perfect date night. First, on Eventbrite, find an art exhibition or jazz concert near Downtown/The Loop. Prioritize options that end before 7:00 PM on Saturday, and check whether the event details mention a dress code or whether it’s suitable for couples. If the current listings do not show a clear end time, choose an event that starts earlier (e.g., before early evening) and has more complete page information, and note in the result whether the end time is clearly provided.945 946After selecting the event, based on its end time (or your noted timing judgment if the end time is unclear) and exact address, go to OpenTable and find a restaurant within walking distance (or a very short drive). I’d like Italian or French cuisine, rated above 4.5, with a price range of $$ or $$$. Most importantly, check whether a table for two is available around 30 minutes after the event ends.947 948Finally, tell me the name of the event, what time it ends (or whether the end time is unknown), and the name of the restaurant you chose, along with the most suitable reservation time."
949task-7fa560,"I’m considering buying clothing from sustainable fashion brands and want to verify whether their environmental claims are credible.950 951First, please search for sustainability-related women’s clothing items on Target (prioritize the keyword **“sustainable”**; if there is no explicit “sustainable” label on the site, use related keywords/tags such as **“organic cotton”** or **“recycled materials”**). Try to find products from **5 different brands**, each with a rating of **4 stars or above**. Record the following for each item: **brand name, product name, price, rating, and the environmental claim mentioned on the product page** (e.g., “made with organic cotton,” “reduced carbon emissions,” etc.). If fewer than 5 qualifying items are found, record the actual number found and specify which criteria were not fully met.952 953Then, go to **Our World in Data** to look up data on **carbon emissions** and **water consumption** in the textile industry. Identify the **global average benchmark values** for textile carbon emissions intensity and water consumption, review the trends shown in the charts, and record key figures.954 955Next, search Etsy for **“sustainable handmade clothing women”** and try to find **3 independent designer shops** with ratings of **4.5 stars or above**. Check how their product descriptions and shop About pages describe their sustainability practices. Record: **shop name, representative product, price, and sustainability-practice description**. If fewer than 3 such shops are found, record the actual number and supplement with relevant shops rated **4.0 stars or above**, clearly noting the rating differences.956 957Finally, output:958- For each Target brand: **brand name, product name, price, rating, sustainability claim description, product link**959- From Our World in Data: **textile industry carbon-emissions and water-consumption benchmark values, plus chart links**960- For each Etsy shop: **shop name, representative product, price, rating, sustainability-practice description, product link**"
961task-7fbc0c,"I’m working on a data visualization project and want to find a solid Python library. Please search Stack Overflow for **“python data visualization”** and check what libraries are recommended in highly upvoted answers. Record the top three answers with the highest vote counts.962 963Then look up each of those recommended libraries on GitHub and review their **star counts**, **whether they’ve been updated recently**, and **how clear and complete their README is**. Help me identify which one is the most active.964 965Finally, search Bilibili for tutorials on that library. Prioritize results by **view count (highest to lowest)**, and find several in-depth tutorials with **high views** and **duration longer than 15 minutes**. Check audience feedback in the comments. If stable sorting by view count is not available, select from the top relevant results and clearly state your selection criteria.966 967At the end, tell me whether this library is worth learning and how good the tutorial quality is."
968task-80d6f4,"I’m very interested in Samsung’s foldable-phone hinge technology and want to find some in-depth technical material. Please search Google Patents for patents filed by Samsung Electronics related to “foldable hinge.” Make sure to use filters so that only **granted patents (Grant)** are shown, and the **filing date is after January 1, 2021**. After finding one that appears to be a core patent, record its patent number and title.  969 970If there are no results under these strict conditions, keep **Grant** unchanged and gradually relax the date condition to **after January 1, 2019**, and specify the actual date range used in your results.  971 972Then go to YouTube and search for a “Samsung Galaxy Z Fold 5 teardown” video that uses this type of technology. Find a popular video with **over 500,000 views**. Open it and verify whether it supports subtitles (CC). If there are no results over 500,000 views, select the video with the highest view count in the current results and note the actual view count.  973 974Finally, tell me the patent number, the video title, and whether subtitles are available."
975task-818fa1,"I am planning a trip to Italy in May 2026, themed around visiting filming locations from the movie ""Mission: Impossible - Dead Reckoning Part One"". First, go to IMDb to find this 2023 movie and identify its main filming cities. From these, find two famous Italian tourist cities (Hint: one is the capital, and one is a 'water city').976 977For the capital city (Rome), search on Google Flights for a one-way ticket from New York (JFK). The planned departure date is May 15, 2026. Use the ""Date Grid"" feature to check prices for surrounding dates. If May 14th or 16th is more than $50 cheaper than May 15th, select the cheaper date; otherwise, keep May 15th. Filter for the *only* non-stop flight available.978 979For the 'water city' (Venice), find a hotel on Tripadvisor. Requirements: It must have the ""Travelers' Choice"" badge, be a 4-star hotel, and have a nightly rate between $300-$600. Sort by ""Traveler Ranked"" and select the top-ranked hotel.980 981Output: The two city names found on IMDb, the selected departure date from Google Flights, the lowest non-stop flight price, the airline, the selected hotel name from Tripadvisor, its nightly rate, whether it has the Travelers' Choice badge, and the respective detail page links for all three websites."
982task-819071,"I’m considering buying a few recently released film noir restoration Blu-rays from the Criterion Collection. First, check the Criterion website for film noir titles added in the last 3 months, filter for 4K UHD or Blu-ray formats, and try to identify 5 films (if fewer than 5 are available, record the actual count). Then, for each film, check IMDb for its rating and number of ratings, prioritizing titles with a rating above 7.5 and more than 5,000 ratings. Next, check Letterboxd for each film’s community rating and number of ratings, prioritizing titles with a rating above 4.0 and more than 2,000 ratings. If fewer than 3 films remain after applying both platform thresholds, fill the shortlist from the collected films by prioritizing how many criteria each title meets (IMDb rating/count and Letterboxd rating/count; the more conditions met, the higher the priority), and clearly mark in the output whether each condition is met along with the corresponding values.983 984Finally, search YouTube for restoration comparison videos for each film (search: “Film Title restoration comparison”), and select comparison videos longer than 3 minutes with relatively high view counts, explaining why each was chosen. Based on restoration specs, cross-platform ratings, and comparison videos, recommend the 3 most worthwhile films to buy (if fewer than 3 candidates are available, recommend the actual number and explain why).985 986For each recommended film, output:987- Title  988- Director  989- Criterion restoration format (4K UHD/Blu-ray)  990- Criterion price  991- Criterion detail page link  992- IMDb rating  993- IMDb rating count  994- IMDb link  995- Letterboxd rating  996- Letterboxd rating count  997- Letterboxd link  998- YouTube comparison video title  999- YouTube view count  1000- YouTube video link  1001- Recommendation rationale (combining restoration quality, ratings, and community reception)"
1002task-82d672,"During a Steam sale event, I'd like to find some engaging RPGs to play during my leisure time.1003 1004First, on Steam, filter for RPG games that are currently on sale (Special Offers), have a user review rating of 'Very Positive' or higher, and are priced between $10 and $30. Sort these results by user reviews and identify the top 5 games.1005 1006Next, on HowLongToBeat, find the 'Main + Extra' average completion time for each of these 5 games.1007 1008Finally, using the current discounted price from Steam, calculate the 'cost per hour of gameplay' (Price / Duration) for each game, and on Metacritic, confirm the Metascore rating for these 5 games.1009 1010For each of the 5 games, output the following: Game Title, Steam Current Price, Steam Review Status, HowLongToBeat Duration, Calculated Cost Per Hour, Metacritic Score, and links to their respective platform detail pages."
1011task-87f4ca,"Our project uses the Python library **requests**, but one of its dependencies had a security incident in the past, so I want to conduct a thorough review this time.1012 1013Please first check **PyPI** for any known vulnerability records related to `requests`, especially those with high severity. Then review its **GitHub repository**, focusing on commit activity over the **past six months** and whether there are any unresolved Issues tagged with **security** (please check multiple pages, not just the first one).1014 1015Finally, search **Google Scholar** for academic papers specifically studying the security of the `requests` library, especially those published in the **last two to three years**, and see whether any have identified architecture-level security risks. If there is no direct research on `requests`, then find studies on the security of Python HTTP client libraries as supplementary evidence and label them as **indirect evidence**.1016 1017Please summarize the findings for me:1018- Whether there are any severe vulnerabilities  1019- Whether maintenance is active  1020- Whether academia has identified potential risks"
1021task-883008,"I’m a die-hard San Francisco Giants fan and I’m planning a pilgrimage to Oracle Park during the next ticket-available part of the season.  1022Please first check a **rolling 15-day window** in that period and identify which Giants home games are available. I want to attend a game against the **Los Angeles Dodgers** or **New York Mets** if possible (prefer a weekend game). If there is no game against those two teams, pick any **weekend home game**. If no suitable home game exists in the current 15-day window, move forward to the **next 15-day window within the same season** and continue searching.1023 1024After confirming the game date, use **Oracle Park as the center point** and find a **3-star or higher hotel within a 15-minute walk** on Google Maps. Then check that hotel on TripAdvisor, filter reviews mentioning **“safety/safe”** or **“noise/quiet,”** and determine whether the sentiment is positive. Finally, find a **well-rated (4.0+ stars) seafood or American restaurant within a 10-minute walk** of the ballpark for a post-game celebration.1025 1026Output required:1027- Game date and opponent  1028- Ballpark location  1029- Hotel name and star rating  1030- Walking time from hotel to ballpark  1031- Summary of TripAdvisor safety/noise review findings  1032- Recommended restaurant name and rating  1033- Verification links for each step"
1034task-8acc4f,"I came across a survey paper on 3D Gaussian Splatting on arXiv. The title is roughly **“A Survey on 3D Gaussian Splatting”**, and the authors seem to be Guo et al. I want to find the latest code implementations mentioned in or related to this survey.1035 1036First, help me confirm the paper’s exact title and arXiv ID on arXiv. Then search GitHub using the paper title, find related repositories, and prioritize them by Star count (highest to lowest), while also checking whether they have been updated within the last three months. If any repository satisfies both **“high stars + updated in the last three months,”** return the one with the highest Star count among them.1037 1038Next, check whether that repository’s README includes a **“Requirements”** or **“Installation”** section, and confirm that it is PyTorch-based.1039 1040Finally, search Hugging Face to see whether there is a Papers page with the same title.  1041- If yes, identify the linked Model with the highest download count.  1042- If no such Papers page exists, search Models using the paper title keywords, pick a highly downloaded relevant model, and state the matching rationale.1043 1044Please compile and return the **arXiv ID**, **GitHub repository link**, and the **most popular model ID on Hugging Face**."
1045task-8b2d27,"I’m looking for a few “hidden gem” films worth collecting—specifically cult movies released by the Criterion Collection that are underrated by critics but beloved by fans.  1046Please first browse the in-stock Blu-ray titles on the Criterion website and try filtering candidates by genre, preferably Horror or Sci-Fi. If there is no usable genre filter at the moment, the filter is malfunctioning, or the filtered results are too limited, continue by manually selecting clearly Horror/Sci-Fi titles (Thriller can be included as a supplement) from the visible in-stock Blu-rays on that page, and state the basis for your classification.1047 1048You need to check each film’s Rotten Tomatoes critic score (Tomatometer) and Letterboxd audience rating. Please try to find 3 films that meet both of these criteria: **Tomatometer below 60%** and **Letterboxd rating above 3.6**.  1049If fewer than 3 films meet the strict criteria, keep **Tomatometer < 60%** as a fixed condition, then fill the remaining slots by selecting titles with the highest Letterboxd ratings, and clearly label which entries strictly meet both conditions and which are supplemental picks, along with their corresponding scores.1050 1051Finally, provide the three films’ titles, directors, exact scores from both platforms, and confirm whether each title is currently marked **In Stock** on the Criterion website."
1052task-8b9af7,"I just moved to Texas and live in San Antonio, TX, but my workplace is in Austin, TX, so I need to commute daily. Since I don’t want to stop for gas too often, I want to buy a used compact SUV with strong real-world range.1053 1054Please first go to Edmunds and find 3 listings that meet all of the following criteria: model year 2022 or newer, under 50,000 miles, priced below $28,000, and currently for sale. For these 3 models, go to Car and Driver and find their corresponding professional reviews (if the exact model year is unavailable, use a nearby year from the same generation).1055 1056Prioritize extracting the “75-mph Highway Driving” fuel-economy result and fuel-tank capacity from the “C/D TEST RESULTS” table (do not use EPA official figures). If a model truly does not have corresponding tested data, replace it with another vehicle that meets the Edmunds criteria, and continue until you have 3 vehicles with available tested data. If repeated attempts still produce fewer than 3, clearly note the shortfall and perform calculations only for the vehicles with available tested data.1057 1058Finally, use Google Maps to plan a driving route from San Antonio to Austin. Based on the round-trip distance, calculate whether each of the 3 vehicles can support my 5-day workweek commute on a full tank under highway-only real-world conditions, without refueling midweek (i.e., fuel tank capacity × tested highway MPG > daily round-trip distance × 5).1059 1060Output: decision recommendation, Edmunds link for each vehicle, Car and Driver review link, tested highway MPG, fuel-tank capacity, theoretical highway range, and whether it meets my one-week commuting requirement."
1061task-8dcd5e,"I’m an event promoter looking for promising indie bands. First, go to **Eventbrite**, search for **Music** events in **Los Angeles**, and use the date filter to show only events happening **this weekend**. Prioritize finding three indie band headline shows that are **not Sold Out** (avoid multi-artist festivals or DJ parties). If fewer than three are available, record the actual number found and continue. Note the name of your top-priority band choice.1062 1063Take that band name to **Spotify** to verify popularity. Open the artist profile page, switch to the **Albums** or **Discography** tab, and check whether they have released any album with “Live” in the title (output **Yes/No**). Then check the **Monthly Listeners** on the main page, which must be above **10,000**. If it is below 10,000, return to Eventbrite and choose the next band until one meets the requirement. If all available candidates in this round are below 10,000, state that in the result and stop. Once a band qualifies, record the #1 song in their **Popular** list.1064 1065Finally, go to **YouTube** to validate live performance strength. Search for a live version of that top song (keyword: **“Song Name Live”**), and use filters to show only videos with duration **Over 4 minutes**. Return up to 5 candidate results (or all if fewer), sort by view count from highest to lowest, and select the top one as the final result. Also include the titles and view counts of the #2 and #3 videos for comparison (if fewer than 3 exist, provide the actual number available).1066 1067**Output:** final selected band name, Eventbrite event link, Spotify monthly listeners, whether they have a Live album, Spotify Top 1 song name, YouTube video title, duration, view count, and link; plus the YouTube #2/#3 titles and view counts (if available)."
1068task-91a67b,"My family and I live in Dallas, Texas. We’ve heard that flu season has arrived and want to strengthen our family’s immunity through diet.1069 1070First, go to the CDC website and check the current flu activity level in Texas (Minimal/Low/Moderate/High/Very High), along with any CDC nutrition-related prevention recommendations. (If the page does not provide explicit nutrition advice, record CDC’s general flu-prevention health recommendations and cite the source.)1071 1072Then, go to WebMD and find an article about immunity-boosting foods. Extract at least 5 immunity-supporting ingredients recommended by CDC or WebMD (for example: citrus fruits, ginger, garlic, spinach).1073 1074Next, use these ingredients as keywords to search for healthy recipes on Food Network. Try to find 5 recipes, prioritizing the following criteria:1075- Rating of 4 stars or above  1076- Prep time within 45 minutes  1077- Each recipe includes at least 2 of the extracted recommended ingredients  1078 1079If fewer than 5 recipes fully meet all criteria, keep “at least 2 recommended ingredients per recipe” as a hard requirement, and relax constraints in this order:10801) rating, then  10812) prep time.  1082 1083In the output, explicitly note for each recipe:1084- its rating  1085- its prep time  1086- whether it meets “rating ≥ 4 stars / time ≤ 45 minutes”.1087 1088Finally, output:1089- Current flu activity level in Texas  1090- Summary of CDC nutrition-related prevention guidance  1091- WebMD article link  1092- List of 5 extracted recommended ingredients  1093- For each recipe: name, rating, prep time, included recommended ingredients, and Food Network recipe detail-page link."
1094task-91ef36,"I am a high school history teacher preparing a lesson on the “impact of the Industrial Revolution on global living standards” (with a focus on the United Kingdom).  1095 1096First, check Wikipedia for “Industrial Revolution” and confirm its approximate start and end years, as well as its place of origin. Then go to Gapminder Tools, select the United Kingdom, and in **Trends** (line chart) mode set the y-axis to **“Income per person (GDP/capita, inflation-adjusted)”**, with time as the default x-axis. Identify the approximate year when the UK’s income per person clearly takes off from **Level 1** (about the $1,000–$2,000 range) and surpasses the **$3,000** threshold, and record the exact value for the most recent complete year.  1097 1098Next, still in Gapminder, switch the y-axis to **“Life expectancy”** and find the UK’s life expectancy values in **1850** and **1950**.  1099 1100Finally, go to Khan Academy, search for “Industrial Revolution,” and find a video or article suitable for an AP history course that includes content related to **“Social impact”** or **“Living standards.”** Confirm the name of the unit it belongs to.  1101 1102Please output: the Industrial Revolution’s start and end years, place of origin, the year the UK’s income per person in Gapminder surpassed $3,000, the UK’s income-per-person value for the most recent complete year, UK life expectancy in 1850 and 1950, the title of the selected Khan Academy content, its unit name, and links to each page."
1103task-925b01,"I’m a food blogger, and for my next family gathering I want to recreate the most authentic Bolognese ragù. There are too many “modernized” versions online, so I want to trace it back to the source.1104 1105First, go to Wikipedia and look up the **“Bolognese sauce”** entry. Review its history and/or any recipe description related to the **Accademia Italiana della Cucina**, and identify two key rules about liquid ingredients:11061. Whether **milk** is required  11072. Whether the traditional wine is **red or white**1108 1109Using these two non-negotiable criteria, search **“Bolognese”** on Allrecipes. Try to find a recipe with a rating above **4.6**, more than **3,000 reviews**, and an ingredient list that strictly matches the two Wikipedia liquid-ingredient requirements.  1110If no recipe meets all conditions at once, keep “ingredient list matches both Wikipedia liquid rules” as the hard requirement, then choose the one with the highest possible rating and review count, and explicitly note how far it falls short of the **4.6 / 3000** threshold.1111 1112Finally, go to YouTube and find a matching hands-on cooking video, with **over 500,000 views** and a duration of **more than 5 minutes** to ensure sufficient detail.1113 1114Output required:  1115- Wikipedia-confirmed milk requirement and wine color  1116- Allrecipes recipe title, rating, and review count  1117- YouTube video title, view count, and duration  1118- Links to each corresponding detail page"
1119task-933495,"I’m researching the medicinal value of **Ginkgo** for improving cognitive function and want to conduct a complete literature review from botany to clinical application.  1120 1121First, go to the **Kew Plants database** to confirm Ginkgo’s accepted scientific name, taxonomic status, and synonyms (to avoid confusion with related species), and record the scientific name, family/genus, and native distribution.  1122 1123Then, use this scientific name to search **PubMed** for pharmacological studies on cognitive function or neuroprotection. Filter for **systematic reviews or meta-analyses** published since **2020** with **free full text**, and try to find **3 highly cited papers**. Record each paper’s title, authors, PMID, and main conclusions. If fewer than 3 are available, supplement with review articles on the same topic and indicate whether each is a systematic review/meta-analysis.  1124 1125Next, go to **Mayo Clinic** to check clinical application information for Ginkgo as an herbal supplement, including the level of evidence for cognitive improvement, as well as side effects and medication warnings.  1126 1127Finally, use **Drugs.com** to check interactions between Ginkgo and common prescription drugs (such as warfarin and aspirin), and record interactions rated **Major** or **Moderate**, along with their explanations.  1128 1129**Output required:**  1130- Ginkgo scientific name, family/genus, Kew page link, and native distribution.  1131- For 3 PubMed papers (or the actual number available): title, authors, PMID, publication year, main conclusions, PubMed link, and whether each is free full text and whether it is a systematic review/meta-analysis.  1132- Mayo Clinic’s evidence-level assessment for Ginkgo in cognitive improvement, side-effect list, and Mayo page link.  1133- From Drugs.com interaction results: drug names with Major/Moderate interactions, severity level, interaction description, and interaction page link."
1134task-957d1a,"I’m transitioning from indie game development to a role at a major studio and plan to apply to game companies in Seattle. First, go to LinkedIn and search for Game Programmer or Engine Engineer positions in Seattle. Filter for jobs posted within the last 30 days, full-time, and Easy Apply. Identify 5 roles whose skill requirements explicitly mention Unreal Engine or Unity, and note the company names. If login requirements prevent access to full job details, record that limitation and continue completing the remaining steps with the visible information.1135 1136Then, go to Glassdoor and check reviews and compensation for each of those 5 companies. Prioritize companies with an overall rating of 3.8 or higher, and find the salary range for the Game Programmer role (base salary, bonus, and total compensation) as well as the work-life balance rating. If ratings or salary details are not visible due to login restrictions, record “Restricted” and keep the company in scope rather than skipping it.1137 1138Next, search these 5 company names on GitHub. For each company, find the open-source project with the highest star count, and review the README to identify the tech stack they use (programming languages, engines, frameworks).1139 1140Finally, based on tech-stack keywords from those projects (e.g., Unreal Engine, C++, Multiplayer Networking), find matching high-quality courses on Udemy. Requirements: rating of at least 4.5, at least 1,000 students, and at least 6 hours of content. Review the course curriculum to confirm whether it includes hands-on projects.1141 1142Output for each company:1143- Company name  1144- LinkedIn job posting link  1145- Job skill requirements list  1146- Glassdoor overall rating  1147- Work-life balance rating  1148- Game Programmer salary range (base salary / bonus / total compensation)  1149- Glassdoor company page link  1150- GitHub open-source project name and star count  1151- Project tech stack  1152- GitHub project link  1153- Recommended Udemy course name  1154- Course rating  1155- Number of students  1156- Course duration  1157- Number of curriculum sections  1158- Udemy course link"
1159task-978ccb,"I have a 65-year-old patient who is currently taking 7 prescription medications and 3 supplements simultaneously. The specific list includes: Prescription medications: Aspirin, Warfarin, Simvastatin, Amlodipine, Metformin, Lisinopril, Omeprazole; Supplements: St. John's Wort, Fish Oil, CoQ10. I need to conduct a comprehensive safety assessment of this medication and supplement regimen.1160 1161First, go to Drugs.com to batch-check the interactions of these 10 medications/supplements. Record all Major and Moderate level interaction combinations along with their specific risk descriptions.1162 1163Next, go to WebMD to individually verify the top 5 high-risk combinations (prioritizing by severity). Confirm if the risk level and clinical significance are consistent.1164 1165Then, search PubMed for clinical evidence regarding these high-risk combinations. Use generic drug names in combination for the search (e.g., ""Warfarin AND St. John's Wort AND interaction""). Filter for clinical trials or case reports published within the last 5 years, finding at least 3 most relevant articles.1166 1167Finally, consult Mayo Clinic's medication guidelines for elderly patients, specifically focusing on special considerations and dosage adjustment recommendations for Warfarin and Simvastatin in the geriatric population.1168 1169**Output:**1170(1) A list of all Major and Moderate level interactions detected by Drugs.com, including the drug combination, severity label, risk description, and a link to the Drugs.com interaction report page.1171(2) The top 5 high-risk combinations verified by WebMD, including the drug combination, the risk level displayed by WebMD, whether it is consistent with Drugs.com, and a link to the WebMD interaction checker result page.1172(3) Three clinical evidence articles found on PubMed, including title, author(s), publication year, journal name, PMID, link to the article details page, and key findings from the abstract.1173(4) Mayo Clinic's medication recommendations for elderly patients, including specific precautions and dosage adjustment suggestions for Warfarin and Simvastatin respectively in the geriatric population, and a link to the corresponding medication page on Mayo Clinic."
1174task-9a23a2,"My father has been diagnosed with early-onset Alzheimer’s disease, and we need to visit a neurology specialist at Mayo Clinic (Rochester campus). First, please find a neurologist on the Mayo Clinic website who specializes in “Memory Disorders” or “Alzheimer’s.” After identifying a doctor, check this doctor’s rating on US News Health to confirm they have a strong reputation.  1175 1176Finally, we need to stay in Rochester for about one month (assume check-in is on the 1st of next month). Please find 3 Airbnb listings. Because my father has limited mobility and we need to cook for ourselves, be sure to filter for properties with both “Elevator” and “Kitchen.” If fewer than 3 listings are available after applying both filters, keep the filters and report the actual number found.  1177 1178Please provide the doctor’s name, US News rating, and the names and total prices of the 3 listings (or the actual number found). Note: We will ultimately need to call the doctor’s office to schedule an appointment, so prioritize doctors whose Mayo profile shows a phone number or an “Appointments/Request appointment” entry point. If the selected doctor’s page does not display a phone number, choose another doctor in the same campus and specialty direction."
1179task-9e1250,"I'm currently researching science-backed weight loss methods and want to develop an actionable diet plan based on academic evidence.1180 1181First, go to Semantic Scholar and search for papers related to 'intermittent fasting'. Filter for papers published since 2023, with over 150 citations, and available with PDF. Find 3 papers, sorted by citation count in descending order. Record the paper title, publication year, citation count, and core fasting method (e.g., 16:8, 5:2, etc.).1182 1183Next, go to WebMD or Mayo Clinic and search for 'intermittent fasting'. Review the clinical recommendations and safety considerations from these medical institutions regarding these fasting methods. Confirm if the methods mentioned in the papers are recognized by mainstream medical organizations. Record the recommended fasting windows and contraindications.1184 1185Then, go to AllRecipes and search for 'high protein low calorie'. Filter recipes with a preparation time under 30 minutes and calories between 300-500 kcal. Find 3 quick recipes suitable for the fasting window.1186 1187Finally, go to Food Network and search for 'healthy dinner recipes'. Filter for recipes with a 5-star rating and under 400 kcal. Find 2 nutrient-dense dinner options. Expand the nutritional information to check if the protein content exceeds 25 grams.1188 1189**Output:**1190For each paper: title, publication year, citation count, Semantic Scholar link, and extracted fasting method.1191For WebMD/Mayo Clinic: a summary of clinical recommendations for the method, safety considerations, recommended fasting windows, and the medical website link.1192For AllRecipes: 3 recipe names, preparation time, calories, and recipe link.1193For Food Network: 2 recipe names, calories, protein content, rating, and recipe link."
1194task-a1d501,"I am currently preparing my final thesis on global extreme poverty (as of December 2025).1195 1196First, please visit OurWorldInData and search for 'Share of population living in extreme poverty'. On the chart page, in the table view, identify the country with the highest poverty rate in the most recent year's data, and record its name, the year, and the specific proportion.1197 1198Next, to gain a deeper understanding of the underlying economic principles, find a 'Development Economics' course on Coursera. Requirements: suitable for 'Intermediate' level, taught in English, and with a rating of 4.7 or higher. Please check the syllabus to confirm if it includes sections on 'Randomized Control Trials' (RCT) or 'Impact Evaluation'.1199 1200Finally, I plan to use R language to process this type of panel data. Go to StackOverflow and search for questions about 'ggplot2 panel data visualization', filtering for the one with the highest number of votes and an accepted answer.

Showing the first 1,200 of 1913 lines. Download the file for the rest.