datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
llm-fol-reasoning-eval
LLM FOL Reasoning Eval
This dataset is derived from ProverQA, a First-Order Logic reasoning benchmark designed to test the ability of large language models (LLMs) to perform structured logical reasoning.It restructures and normalizes the ProverQA development and training data into a unified, clean format suitable for evaluating chain-of-thought (CoT) and symbolic reasoning capabilities in LLMs.
Source
Original dataset: ProverQA: A First-Order Logic Reasoning… See the full description on the dataset page: https://huggingface.co/datasets/MinaGabriel/llm-fol-reasoning-eval.twi-llm-reasoning-dataset-1k
This dataset is shared under CC BY-NC 4.0, which means you are free to use, share, and adapt it for non-commercial research and educational purposes with attribution. You can read the full license at https://creativecommons.org/licenses/by-nc/4.0/.
This dataset is made available because of Ghana NLP's volunteer driven research work. Please consider contributing to any of our projects on Github
Twi Reasoning Dataset
A Twi (Akan) translation of the Multilingual-Thinking… See the full description on the dataset page: https://huggingface.co/datasets/ghanaopenai/twi-llm-reasoning-dataset-1k.chem-reasoning-llmllm-ground-truth-reasoningLiveMathBench-en-with-reasoningchess-sft-20k-llm-reasoning-enriched-dpo-hard-negatives-v1llm_aggrefact_pre_aug9_no_reasoningft-llm-2026-reasoning-sft
FT-LLM 2026 Reasoning SFT
A supervised fine-tuning (SFT) dataset of mathematical chain-of-thought traces used in Phase 2 of the COMPASS project to enhance the reasoning capability of a Japanese VLM's LLM backbone. Reasoning traces were distilled from a Qwen3-30B teacher over a broad pool of public math-reasoning datasets (GSM8K, MATH, SVAMP, AQuA-RAT, MathInstruct, MGSM-ja, Orca Math, NuminaMath-CoT, OpenR1-Math, Open Math Reasoning, and others). Each sample follows a structured XML… See the full description on the dataset page: https://huggingface.co/datasets/Yana/ft-llm-2026-reasoning-sft.ft-llm-2026-reasoning-dpo
FT-LLM 2026 Reasoning DPO
A Direct Preference Optimization (DPO) dataset of (prompt, chosen, rejected) triples used in Phase 2 of the COMPASS project to align a Japanese VLM's LLM backbone toward correct mathematical reasoning. The chosen responses are chain-of-thought traces distilled from a Qwen3-30B teacher in the structured <Problem>/<Thinking>/<Answer> XML format. The rejected responses are synthetically generated by corrupting the chosen responses under three strategies, mixed… See the full description on the dataset page: https://huggingface.co/datasets/Yana/ft-llm-2026-reasoning-dpo.hayabusa_llm_report_forensic_reasoningOmni-MATH-imo-imc-with-reasoning2OlymMATH-en-with-reasoningllm-math-reasoning-dataset株式会社APTOは、大規模言語モデル（LLM）の数理推論能力向上を目指し、データセットの開発を行いました。
近年、LLMの性能は飛躍的に向上していますが、複数ステップの計算や厳密な解答形式を要求される数理タスクでは、依然として誤答や形式不備が散見されます。
これらの課題に対応するため、数理推論の思考・回答精度を高めることを目的とし、LLM学習用データセットを開発しました。
LLMを取り扱っている方々は、以下のような数学応用における課題を経験されたことがあると思います。
段階的な計算を展開しない、計算プロセスを正確に踏まず解答を誤る
整数や分数など、指定された解答フォーマットに従わない出力
途中式の記載漏れや最終解答のみの出力など、問題へのアプローチが見えない解答
このように複雑な数理問題の回答において、指示やルールを無視し正確なアウトプットが行われないという経験をされた方も多いと思います。
そこで数学問題の解答精度を改善するために、LLMの思考能力改善のノウハウを活かし、複雑な思考プロセスを含む数学問題データセットの開発を行いました。
データの内容… See the full description on the dataset page: https://huggingface.co/datasets/APTO-001/llm-math-reasoning-dataset.team-truthowl-mixed-reasoning-dataset
Team P11 Mixed Reasoning Dataset
📊 Dataset description
HLE（Humanity's Last Exam）向けに作成した、数学中心＋科学MCの混合推論データセットです。
推論過程（Chain-of-Thought）を保持し、最終解答の正規化を行っています。
対象モデルは DeepSeek-R1-Distill-Qwen-32B、学習はQLoRAを想定しています。
🎯 Purpose
Competition: 松尾研LLMコンペ 2025
Target Model: DeepSeek-R1-Distill-Qwen-32B
Training Method: QLoRA Fine-tuning（4bit NF4, double quant）
📦 Composition
Math Hard（MATH Level≥3, HARDMath）
Math Mid（GSM8K, MetaMathQA）
Science（GPQA… See the full description on the dataset page: https://huggingface.co/datasets/weblab-llm-competition-2025-bridge/team-truthowl-mixed-reasoning-dataset.Nemotron-CrossThink-QA_reasoning_Phi-4-reasoning-plus_0803_n20480_test_5kllm-complex-reasoning-train-qwen2-72b-instruct-correctu-math-with-reasoningreasoning-llmchess-sft-20k-llm-reasoning-enrichedcpt112_and_llm_split_reasoning_sftcpt112_and_llm_split_reasoning_sft_recallHiST-LLM_added_hist_reasoning_v2team-camino-Nemotron-CrossThink-QA_reasoning_Phi-4-reasoning-plus_0803_n20480_test_5kchess-sft-2k-llm-reasoning-enrichedreasoning-llm-entity-training-ver-1.2llm-ground-truth-reasoning-extendedEAS-Qwen3-30B-A3B-Thinking-2507-cd8arg-iterative-llm-extract-test
EAS-Qwen3-30B-A3B-Thinking-2507-cd8arg-iterative-llm-extract-test
LLM-based math span extraction with canonicalization
Dataset Info
Rows: 1
Columns: 26
Columns
Column
Type
Description
question
Value('string')
No description provided
metadata
Value('string')
No description provided
task_source
Value('string')
No description provided
formatted_prompt
List({'content': Value('string'), 'role': Value('string')})
No description provided… See the full description on the dataset page: https://huggingface.co/datasets/reasoning-degeneration-dev/EAS-Qwen3-30B-A3B-Thinking-2507-cd8arg-iterative-llm-extract-test.reasoning-llm-entity-training-ver-1.1reasoning-llm-entity-training-ver-2.0EAS-Qwen3-30B-A3B-Thinking-2507-cd8arg-iterative-llm-extract-test-stage02a
EAS-Qwen3-30B-A3B-Thinking-2507-cd8arg-iterative-llm-extract-test-stage02a
Deterministic extraction of arithmetic attempts, expressions, subgoals, branches, and failure markers from Countdown reasoning traces
Dataset Info
Rows: 1
Columns: 26
Columns
Column
Type
Description
question
Value('string')
No description provided
metadata
Value('string')
No description provided
task_source
Value('string')
No description provided
formatted_prompt… See the full description on the dataset page: https://huggingface.co/datasets/reasoning-degeneration-dev/EAS-Qwen3-30B-A3B-Thinking-2507-cd8arg-iterative-llm-extract-test-stage02a.