datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
IFEval
Dataset Card for IFEval
Dataset Summary
This dataset contains the prompts used in the Instruction-Following Eval (IFEval) benchmark for large language models. It contains around 500 "verifiable instructions" such as "write in more than 400 words" and "mention the keyword of AI at least 3 times" which can be verified by heuristics. To load the dataset, run:
from datasets import load_dataset
ifeval = load_dataset("google/IFEval")
Supported Tasks and… See the full description on the dataset page: https://huggingface.co/datasets/google/IFEval.bigbenchThe Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to
probe large language models, and extrapolate their future capabilities.google-code-archive
Google Code Archive Dataset
Dataset Description
This dataset was compiled from the Google Code Archive, a preserved snapshot of projects hosted on Google Code, Google's open-source project hosting service that operated from 2006 to 2016. Google Code was one of the major code hosting platforms of its era, hosting hundreds of thousands of open-source projects before its shutdown. The archive provides a unique historical record of open-source development during a formative… See the full description on the dataset page: https://huggingface.co/datasets/nyuuzyou/google-code-archive.taskmaster2Taskmaster is dataset for goal oriented conversations. The Taskmaster-2 dataset consists of 17,289 dialogs in the seven domains which include restaurants, food ordering, movies, hotels, flights, music and sports. Unlike Taskmaster-1, which includes both written "self-dialogs" and spoken two-person dialogs, Taskmaster-2 consists entirely of spoken two-person dialogs. In addition, while Taskmaster-1 is almost exclusively task-based, Taskmaster-2 contains a good number of search- and recommendation-oriented dialogs. All dialogs in this release were created using a Wizard of Oz (WOz) methodology in which crowdsourced workers played the role of a 'user' and trained call center operators played the role of the 'assistant'. In this way, users were led to believe they were interacting with an automated system that “spoke” using text-to-speech (TTS) even though it was in fact a human behind the scenes. As a result, users could express themselves however they chose in the context of an automated interface.code_x_glue_cc_code_completion_token
Dataset Card for "code_x_glue_cc_code_completion_token"
Dataset Summary
CodeXGLUE CodeCompletion-token dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/CodeCompletion-token
Predict next code token given context of previous tokens. Models are evaluated by token level accuracy.
Code completion is a one of the most widely used features in software development through IDEs. An effective code completion tool could improve software… See the full description on the dataset page: https://huggingface.co/datasets/google/code_x_glue_cc_code_completion_token.unpredictable_support-google-comThe UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance. For more details please see the accompanying dataset card.schema_guided_dstc8The Schema-Guided Dialogue dataset (SGD) was developed for the Dialogue State Tracking task of the Eights Dialogue Systems Technology Challenge (dstc8).
The SGD dataset consists of over 18k annotated multi-domain, task-oriented conversations between a human and a virtual assistant.
These conversations involve interactions with services and APIs spanning 17 domains, ranging from banks and events to media, calendar, travel, and weather.
For most of these domains, the SGD dataset contains multiple different APIs, many of which have overlapping functionalities but different interfaces,
which reflects common real-world scenarios.google-code-archive
Google Code Archive Dataset
Dataset Description
This dataset was compiled from the Google Code Archive, a preserved snapshot of projects hosted on Google Code, Google's open-source project hosting service that operated from 2006 to 2016. Google Code was one of the major code hosting platforms of its era, hosting hundreds of thousands of open-source projects before its shutdown. The archive provides a unique historical record of open-source development during a formative… See the full description on the dataset page: https://huggingface.co/datasets/Carrillo16/google-code-archive.unpredictable_support-google-comThe UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance. For more details please see the accompanying dataset card.google-code-archive
Google Code Archive Dataset
Dataset Description
This dataset was compiled from the Google Code Archive, a preserved snapshot of projects hosted on Google Code, Google's open-source project hosting service that operated from 2006 to 2016. Google Code was one of the major code hosting platforms of its era, hosting hundreds of thousands of open-source projects before its shutdown. The archive provides a unique historical record of open-source development during a… See the full description on the dataset page: https://huggingface.co/datasets/0xzanuee/google-code-archive.google-code-archive
Google Code Archive Dataset
Dataset Description
This dataset was compiled from the Google Code Archive, a preserved snapshot of projects hosted on Google Code, Google's open-source project hosting service that operated from 2006 to 2016. Google Code was one of the major code hosting platforms of its era, hosting hundreds of thousands of open-source projects before its shutdown. The archive provides a unique historical record of open-source development during a formative… See the full description on the dataset page: https://huggingface.co/datasets/Mgmgrand420/google-code-archive.code_x_glue_cc_code_completion_line
Dataset Card for "code_x_glue_cc_code_completion_line"
Dataset Summary
CodeXGLUE CodeCompletion-line dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/CodeCompletion-line
Complete the unfinished line given previous context. Models are evaluated by exact match and edit similarity.
We propose line completion task to test model's ability to autocomplete a line. Majority code completion systems behave well in token level completion, but fail in… See the full description on the dataset page: https://huggingface.co/datasets/google/code_x_glue_cc_code_completion_line.air_dialogue
Dataset Card for air_dialogue
Dataset Summary
AirDialogue, is a large dataset that contains 402,038 goal-oriented conversations. To collect this dataset, we create a contextgenerator which provides travel and flight restrictions. Then the human annotators are asked to play the role of a customer or an agent and interact with the goal of successfully booking a trip given the restrictions.
News in v1.3:
We have included the test split of the AirDialogue dataset.
We have… See the full description on the dataset page: https://huggingface.co/datasets/google/air_dialogue.code_x_glue_cc_cloze_testing_all
Dataset Card for "code_x_glue_cc_cloze_testing_all"
Dataset Summary
CodeXGLUE ClozeTesting-all dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/ClozeTesting-all
Cloze tests are widely adopted in Natural Languages Processing to evaluate the performance of the trained language models. The task is aimed to predict the answers for the blank with the context of the blank, which can be formulated as a multi-choice classification problem.
Here we… See the full description on the dataset page: https://huggingface.co/datasets/google/code_x_glue_cc_cloze_testing_all.taskmaster1Taskmaster-1 is a goal-oriented conversational dataset. It includes 13,215 task-based dialogs comprising six domains. Two procedures were used to create this collection, each with unique advantages. The first involves a two-person, spoken "Wizard of Oz" (WOz) approach in which trained agents and crowdsourced workers interact to complete the task while the second is "self-dialog" in which crowdsourced workers write the entire dialog themselves.code_x_glue_cc_cloze_testing_maxmin
Dataset Card for "code_x_glue_cc_cloze_testing_maxmin"
Dataset Summary
CodeXGLUE ClozeTesting-maxmin dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Code-Code/ClozeTesting-maxmin
Cloze tests are widely adopted in Natural Languages Processing to evaluate the performance of the trained language models. The task is aimed to predict the answers for the blank with the context of the blank, which can be formulated as a multi-choice classification problem.… See the full description on the dataset page: https://huggingface.co/datasets/google/code_x_glue_cc_cloze_testing_maxmin.taskmaster3Taskmaster is dataset for goal oriented conversations. The Taskmaster-3 dataset consists of 23,757 movie ticketing dialogs. By "movie ticketing" we mean conversations where the customer's goal is to purchase tickets after deciding on theater, time, movie name, number of tickets, and date, or opt out of the transaction. This collection was created using the "self-dialog" method. This means a single, crowd-sourced worker is paid to create a conversation writing turns for both speakers, i.e. the customer and the ticketing agent.coached_conv_prefA dataset consisting of 502 English dialogs with 12,000 annotated utterances between a user and an assistant discussing
movie preferences in natural language. It was collected using a Wizard-of-Oz methodology between two paid crowd-workers,
where one worker plays the role of an 'assistant', while the other plays the role of a 'user'. The 'assistant' elicits
the 'user’s' preferences about movies following a Coached Conversational Preference Elicitation (CCPE) method. The
assistant asks questions designed to minimize the bias in the terminology the 'user' employs to convey his or her
preferences as much as possible, and to obtain these preferences in natural language. Each dialog is annotated with
entity mentions, preferences expressed about entities, descriptions of entities provided, and other statements of
entities.task674_google_wellformed_query_sentence_generation
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task674_google_wellformed_query_sentence_generation
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative Instructions on… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task674_google_wellformed_query_sentence_generation.google-code-archive
Google Code Archive Dataset
Dataset Description
This dataset was compiled from the Google Code Archive, a preserved snapshot of projects hosted on Google Code, Google's open-source project hosting service that operated from 2006 to 2016. Google Code was one of the major code hosting platforms of its era, hosting hundreds of thousands of open-source projects before its shutdown. The archive provides a unique historical record of open-source development during a formative… See the full description on the dataset page: https://huggingface.co/datasets/NarsAI/google-code-archive.task675_google_wellformed_query_sentence_generation
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task675_google_wellformed_query_sentence_generation
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative Instructions on… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task675_google_wellformed_query_sentence_generation.rfm-rm-as-user-dataset
RFM Reward Model As User Dataset
This dataset was generated for the NeurIPS 2025 paper titled "Capturing Individual Human Preferences with Reward Features". It is released to support the reproducibility of the experiments described in the paper, particularly those in the "Modelling groups of real users" section.
Instead of containing preferences from human raters, this dataset uses 8 publicly available reward models (RMs) as proxies for human raters. This allows for large-scale… See the full description on the dataset page: https://huggingface.co/datasets/google/rfm-rm-as-user-dataset.google-leak-dataset
Google Leak Dataset
A dataset extracted from the google leak under Apache 2.0 license that occurred on May 27, 2024.
Dataset description:
The stored data has been altered but not modified and corresponds to the titles and descriptions of both modules and the keys and values of those modules, provided by the leaked google content warehouse api documentation version 0.4.0.
Data alteration explanation:
Data has been pruned to get only titles and descriptions.
All… See the full description on the dataset page: https://huggingface.co/datasets/elseodelasgalletas/google-leak-dataset.ai-vs-human-google-gemma-2-2b-it
AI vs Human dataset on the CNN Daily mails
Dataset Description
This dataset showcases pairs of truncated articles and their respective completions, crafted either by humans or an AI language model.
Each article was randomly truncated between 25% and 50% of its length.
The language model was then tasked with generating a completion that mirrored the characters count of the original human-written continuation.
Data Fields
'human': The original human-authored… See the full description on the dataset page: https://huggingface.co/datasets/zcamz/ai-vs-human-google-gemma-2-2b-it.nlp-google-reviews-dataset
NLP Google Reviews Dataset
A curated, multi-source dataset of 516 real Google reviews prepared for NLP tasks such as sentiment analysis, text classification, and topic modelling.
Dataset Description
This dataset was built using a production-grade Python pipeline that collects Google reviews from three independent sources, cleans and normalizes the data, and merges everything into a single structured CSV.
Sources
public_dataset: 495 reviews
web_scraping: 16… See the full description on the dataset page: https://huggingface.co/datasets/talhaa/nlp-google-reviews-dataset.task673_google_wellformed_query_classification
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task673_google_wellformed_query_classification
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task673_google_wellformed_query_classification.python-google-style-docstrings
Overview
This dataset contains Python code-docstring pairs, whereas the docstrings are in Google style. A Google style docstring is structured as follows:
<Description of the code>
Args:
<var1> (<data-type>) : <description of var1>
<var2> (<data_type>) : <description of var2>
Returns:
<var3> (<data-type>) : <description of var3>
Raises:
<var4> (<data-type>) : <description of var4>
The format varies widely (like additional sections such as Examples, Notes, etc) but generally… See the full description on the dataset page: https://huggingface.co/datasets/Mir-2002/python-google-style-docstrings.google-smol-en-ru-kjh
⚠️ Notice: This dataset has moved to google/smol under the subsets smoldoc__ru_kjh and smolsent__ru_kjh.
Please use the official repository instead.
google-smol-en-ru-kjh
This dataset contains the Khakas translation of SmolDoc and SmolSent, as well as the English and Russian versions of
these documents. SMOL (Set for Maximal Overall Leverage) is a collection professional translations into 221 Low-Resource
Languages, for the purpose
of training translation models, and otherwise increasing… See the full description on the dataset page: https://huggingface.co/datasets/adeshkin/google-smol-en-ru-kjh.google-gemma-2-9b-it__llm-quality-factual-mini__019e3b975288
google/gemma-2-9b-it on llm.quality.factual-mini (NVIDIA H100 80GB HBM3)
Back to leaderboard
Headline metrics
Metric
Value
Unit
N Samples
10
N Ok
10
Ok Rate
1
Accuracy
1
Accuracy P05
1
Accuracy P50
1
Accuracy P95
1
TTFT P50
16.8943
ms
Total P50 Ms
142.032
Tokens Out Total
205
Run configuration
Model: google/gemma-2-9b-it @ unknown00
Engine: vllm vunknown
Quantization: fp16
Hardware: NVIDIA H100 80GB HBM3
Driver:… See the full description on the dataset page: https://huggingface.co/datasets/Yobitel/google-gemma-2-9b-it__llm-quality-factual-mini__019e3b975288.google-gemma-2-9b-it__llm-inference-chatbot-short__019e3b973b4f
google/gemma-2-9b-it on llm.inference.chatbot-short (NVIDIA H100 80GB HBM3)
Back to leaderboard
Headline metrics
Metric
Value
Unit
TTFT P50
30.0542
ms
TTFT P99
183.9538
ms
TPOT P50
8.648
ms
TPOT P99
10.7976
ms
Total P50 Ms
1095.4794
Total P99 Ms
1422.5879
Req Per S Passing
3.3476
Req Per S All
3.3947
Compliance Rate
0.9861
Ok Rate
1
Throughput Tok Per S
385.0208
Power Avg W
901.0112
Power Peak W
937.794
Energy Joules… See the full description on the dataset page: https://huggingface.co/datasets/Yobitel/google-gemma-2-9b-it__llm-inference-chatbot-short__019e3b973b4f.