datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
or-bench
OR-Bench: An Over-Refusal Benchmark for Large Language Models
Please see our demo at HuggingFace Spaces.
Overall Plots of Model Performances
Below is the overall model performance. X axis shows the rejection rate on OR-Bench-Hard-1K and Y axis shows the rejection rate on OR-Bench-Toxic. The best aligned model should be on the top left corner of the plot where the model rejects the most number of toxic prompts and least number of safe prompts. We also plot a blue line… See the full description on the dataset page: https://huggingface.co/datasets/bench-llm/or-bench.or-bench
OR-Bench: An Over-Refusal Benchmark for Large Language Models
Please see our demo at HuggingFace Spaces.
Overall Plots of Model Performances
Below is the overall model performance. X axis shows the rejection rate on OR-Bench-Hard-1K and Y axis shows the rejection rate on OR-Bench-Toxic. The best aligned model should be on the top left corner of the plot where the model rejects the most number of toxic prompts and least number of safe prompts. We also plot a blue line… See the full description on the dataset page: https://huggingface.co/datasets/bench-llms/or-bench.OR-Space
OR-Space
A full-lifecycle workspace benchmark for industrial optimization agents.
OR-Space evaluates whether language-model agents can work reliably with
operations research problems represented as executable, multi-file workspaces.
Rather than presenting a self-contained mathematical prompt, each task
distributes evidence across business requirements, structured data, source
code, execution logs, and solver records.
The benchmark contains 100 optimization topologies. Each… See the full description on the dataset page: https://huggingface.co/datasets/Chenyu-Zhou/OR-Space.or-bench
OR-Bench: An Over-Refusal Benchmark for Large Language Models
Please see our leaderboard at HuggingFace Spaces.
Overall Plots of Model Performances
Below is the overall model performance. X axis shows the rejection rate on OR-Bench-Hard-1K and Y axis shows the rejection rate on OR-Bench-Toxic. The best aligned model should be on the top left corner of the plot where the model rejects the most number of toxic prompts and least number of safe prompts. We also plot a blue… See the full description on the dataset page: https://huggingface.co/datasets/orbench-llm/or-bench.Origin-Sequence-Data
AL-GR/Origin-Sequence-Data: Raw User Behavior Sequences 📜
About the Dataset
Each row in this dataset (Origin-Sequence-Data) represents a step in a user's journey, consisting of a sequence of previously interacted items (user_history) and the next item they interacted with (target_item). All item IDs have been anonymized into short, unique strings.
This dataset is ideal for:
🧑🔬 Researchers who want to design their own data processing or prompting strategies for… See the full description on the dataset page: https://huggingface.co/datasets/AL-GR/Origin-Sequence-Data.or-bench-toxic-all
OR-Bench: An Over-Refusal Benchmark for Large Language Models
This dataset constains highly toxic prompts, use with caution!!!
Please see our demo at HuggingFace Spaces.
Overall Plots of Model Performances
Below is the overall model performance. X axis shows the rejection rate on OR-Bench-Hard-1K and Y axis shows the rejection rate on OR-Bench-Toxic. The best aligned model should be on the top left corner of the plot where the model rejects the most number of toxic… See the full description on the dataset page: https://huggingface.co/datasets/bench-llms/or-bench-toxic-all.OracleProto
OracleProto: Forecasting Evaluation Set
Chinese doc: [中文文档]
GitHub repo: [MaYiding/OracleProto]
Visit Our Leaderboards: [Website]
View Our Paper: [arXiv]
A SQLite-packaged evaluation set of 80 hand-curated forecasting questions on real-world events, with resolution dates between 2026-03-12 and 2026-04-14, released alongside the GitHub Repo. Both the rows and the byte-stable prompt-reconstruction recipe are packaged in a single file, forecast_eval_set_example.db, which exposes two… See the full description on the dataset page: https://huggingface.co/datasets/MaYiding/OracleProto.or-bench
OR-Bench: An Over-Refusal Benchmark for Large Language Models
Please see our demo at HuggingFace Spaces.
Overall Plots of Model Performances
Below is the overall model performance. X axis shows the rejection rate on OR-Bench-Hard-1K and Y axis shows the rejection rate on OR-Bench-Toxic. The best aligned model should be on the top left corner of the plot where the model rejects the most number of toxic prompts and least number of safe prompts. We also plot a blue… See the full description on the dataset page: https://huggingface.co/datasets/jerogo/or-bench.SWEbench-Verified-eval150-M2.7-Qwen3.5-9B-orch-7arms-2repeats-w32-20260920
SWE-bench Verified eval150 — M2.7 × Qwen3.5-9B, seven arms, two repeats, 32 concurrency
Campaign 2026-09-20. 14/14 independent full150 runs audited. Complete accuracy evidence.
Evaluation mode is orch: MiniMax-M2.7 orchestrator and the specified Qwen3.5-9B worker. Training mode is labeled independently. All runs use 32 concurrent episodes, 10GiB Docker sandboxes, four TP1 workers and one TP4/EP4 coordinator. Frozen regression-gated prompts, decoding and canonical verifier match… See the full description on the dataset page: https://huggingface.co/datasets/CharlieLLL/SWEbench-Verified-eval150-M2.7-Qwen3.5-9B-orch-7arms-2repeats-w32-20260920.OrigamIM
OrigamIM: An Ambiguous Dataset of Sentence Interpretations, Implicit Moral Judgments and Reader Impressions
Introduction
Please cite following papers when using the origamIM dataset (paper 1 and paper 2):
Allein, Liesbeth, and Marie-Francine Moens. "OrigamIM: An Ambiguous Dataset of Sentence Interpretations, Implicit Moral Judgments and Reader Impressions." Proceedings of the 3rd Workshop on Perspectivist Approaches to NLP @LREC-COLING 2024 (2024).
Allein… See the full description on the dataset page: https://huggingface.co/datasets/laallein/OrigamIM.coachtwin-workouts
CoachTwin Workouts
10,393 synthetic, structured workout plans, generated with an open
small language model and used to power the
CoachTwin app -
a workout recommender plus AI workout generator.
How it was generated
Generated with Qwen2.5-Instruct using parameterised one-shot prompting over a
1,920-cell grid (goal x equipment x duration x difficulty x body_focus),
followed by a deterministic repair-then-validate post-processing pass.
The corpus comes from two runs… See the full description on the dataset page: https://huggingface.co/datasets/OrDora/coachtwin-workouts.tamil-orca
Tamil Orca-Style Dataset
Overview
This repository hosts the Tamil Orca-style dataset, meticulously curated to enhance the reasoning capabilities of large language models in Tamil. The dataset is a fusion of translations and responses generated by GPT-4 and Gemini models.
Content: The dataset contains three columns - 'Instruction', 'Query', and 'Answer'.
Purpose: It's designed to significantly improve the reasoning capability of AI language models in Tamil.
Usage: If… See the full description on the dataset page: https://huggingface.co/datasets/azharmo/tamil-orca.Classical-Mechanics-Equations-Dataset_SFT-or-LoRA
Classical Mechanics Equations Dataset (SFT / LoRA Ready)
A structured dataset of 64 classical mechanics equations from Newtonian,
Lagrangian, and Hamiltonian mechanics, expanded into 448 instruction-tuning
rows across three task types: equation explanation, Q&A, and derivation.
Designed for fine-tuning LLMs on physics reasoning, STEM Q&A, and
equation understanding tasks.
Overview
Property
Value
Domain
Classical Mechanics (Physics)
Total rows
448
Train… See the full description on the dataset page: https://huggingface.co/datasets/ChaoticEconomist/Classical-Mechanics-Equations-Dataset_SFT-or-LoRA.ORCHESTRA-simple-1M
ORCHESTRA-simple-1M
GitHub: nk2028/ORCHESTRA-dataset
中文簡介
ORCHESTRA (cOmpRehensive Classical cHinESe poeTRy dAtaset) 是一個全面的古典中文詩歌的數據集，數據來自搜韻網。本數據集由 nk2028 進行格式轉換並發佈，希望透過公開高品質的古典中文詩歌數據，促進對古典中文詩歌及古典中文自然語言處理的研究。
ORCHESTRA-simple 是 ORCHESTRA 數據集的簡化格式，僅保留 id, title, group_index, type, dynasty, author, content 這 7 個欄位，而去除其他欄位，以簡化使用。
本資料集可用於大型語言模型的訓練。如欲作其他用途，請向數據提供者搜韻網諮詢。
English Introduction
ORCHESTRA (cOmpRehensive Classical cHinESe poeTRy dAtaset) is a comprehensive dataset of classical… See the full description on the dataset page: https://huggingface.co/datasets/Ayaka/ORCHESTRA-simple-1M.en_wikipedia_001
Dataset Card for en_wikipedia_001
The en_wikipedia_001 dataset is a collection of crawled paragraph text from Wikipedia on the 28th of April, 2024. It contains high-quality text, stored in multiple documents, available to be used to finetune or train AI models based that the license is followed.
Dataset Details
The dataset was crawled using our web crawler on the 28th of April at an average of 1 page per second as to respect robots.txt rules. Strict licensing must be… See the full description on the dataset page: https://huggingface.co/datasets/orionai/en_wikipedia_001.OR-Space
OR-Space
A full-lifecycle workspace benchmark for industrial optimization agents.
OR-Space evaluates whether LLM agents can do reliable operations research work
inside executable, multi-file workspaces. Each instance keeps business
requirements, parameter files, source code, solver artifacts, and evaluation
metadata as separate files, forcing the agent to recover and maintain the
optimization model through workspace interaction rather than one-shot text
generation.… See the full description on the dataset page: https://huggingface.co/datasets/YiYao7017/OR-Space.Jazz-Blues-Music-Dataset_SFT-or-LoRA
Jazz & Blues Music Dataset (SFT / LoRA Ready)
A structured dataset covering 82 iconic Jazz and Blues songs, 21 artist
profiles, and 41 historical events, expanded into 1,219
instruction-tuning rows across 7 task types.
Designed for fine-tuning LLMs on music knowledge, cultural history, artist
biography, and domain-specific Q&A tasks.
Overview
Property
Value
Domain
Jazz & Blues Music
Total rows
1,219
Train split
1,036 (85%)
Validation split
91 (~7.5%)… See the full description on the dataset page: https://huggingface.co/datasets/ChaoticEconomist/Jazz-Blues-Music-Dataset_SFT-or-LoRA.tamil-orca-transliterated
Tamil Orca-Style Dataset
Overview
This repository hosts the Tamil Orca-style transliterated dataset, meticulously curated to enhance the reasoning capabilities of large language models in Tamil. The dataset is a transliterated version tamil-orca fusion of translations and responses generated by GPT-4 and Gemini models.
Content: The dataset contains three columns - 'Instruction', 'Query', and 'Answer'.
Purpose: It's designed to significantly improve the reasoning… See the full description on the dataset page: https://huggingface.co/datasets/azharmo/tamil-orca-transliterated.Orin-Character-JP-v1ExplainableAI-emotions-DPO-ORPO-RLHF
Preference Dataset for Explainable Multi-Label Emotion Classification
This repository contains a preference dataset compiled to compare two model-generated responses for explaining multi-label emotion classifications on Tweets. The dataset is accompanied by human annotations indicating which response was preferred, based on a set of defined dimensions (clarity, correctness, helpfulness, and verbosity). The annotation guidelines are included to describe how these preference judgments… See the full description on the dataset page: https://huggingface.co/datasets/imhmdf/ExplainableAI-emotions-DPO-ORPO-RLHF.Roomly-Student-Bios-Multimodal
Roomly: Multimodal Roommate Matching Dataset
🎯 Problem Statement
Finding a roommate is often reduced to dry filters like "budget" and "location". Roomly aims to revolutionize this by focusing on personality, lifestyle, and visual preferences. This dataset provides synthetic student profiles and their ideal room environments.
📊 Exploratory Data Analysis (EDA)
1. User Persona Distribution
Our dataset contains a balanced mix of different student… See the full description on the dataset page: https://huggingface.co/datasets/Orib24/Roomly-Student-Bios-Multimodal.