Medical QAs from Chinese encyclopedia
364,420 Chinese medical question-answer pairs for training and evaluating medical LLMs.
- Text generation
- Chinese
Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.
Answered by Gemini 4 · the chat is kept only in this tab
364,420 Chinese medical question-answer pairs for training and evaluating medical LLMs.
471 real GitHub tasks for evaluating AI agents in writing software patches in less popular languages.
A cleaned English SFT dataset from Claude Fable-5 agent traces, 12,730 records.
English dataset for evaluating models in Visual Question Answering and table comprehension tasks.
4,659 English single-turn examples from the closed Claude Fable-5 preview model for agent SFT training.
Distilled step-by-step reasoning from Qwen3.8-27B across four domains for training thinking models.
Huge multilingual corpus of web texts, code, scientific papers, and mathematics for the first LLM pre-training stage.
1K-10K examples to test if a model follows a given JSON/YAML schema under various prompt formulations.
A specialized benchmark by IBM Research for evaluating LLMs and multi-agent systems in industrial operations.
English question-answer pairs on electrical engineering, KiCad EDA, and Python scripts.
Open dataset of 49k chemistry questions with short answers for training and evaluating models using RL.
234k question-answer pairs for scientific reasoning, synthesized from articles via an argumentation skeleton.
Physics, biology, and chemistry problems from NVIDIA with verifier settings for reinforcement learning.
Over 4.7 million cleaned user dialogues with language models from AllenAI.
242k predominantly GPT-4 examples from open sources for instruction tuning.
An English dataset of 12,000 examples for training and evaluating retrieval-augmented generation models.
Selection of WikiHow articles with structured lists for summarization and question answering tasks.
An English corpus of 100K to 1M examples for translating textual conditions into PDDL planning language.
Manually curated descriptions of over 2,000 Creative Commons licensed YouTube channels.
Digitized U.S. court decisions in English for legal text training and analysis.
A large collection of texts for training language models, including code.
1 million to 10 million Chinese instruction-response pairs for SFT, including DeepSeek-R1 distillation.
10k–100k Chinese instruction-response pairs for text model fine-tuning; research use only.
Set of 21,155 Korean Alpaca-style instructions and responses for training chat models.