{"data":{"slug":"hkust-nlp-dart-math","name":"dart-math","tagline":"[NeurIPS'24] Official code for *🎯DART-Math: Difficulty-Aware Rejection Tuning for Mathematical Problem-Solving*","github_url":"https://github.com/hkust-nlp/dart-math","owner":"hkust-nlp","repo":"dart-math","owner_avatar_url":"https://avatars.githubusercontent.com/u/116073107?v=4","primary_language":"Jupyter Notebook","stars":120,"forks":8,"topics":["deep-learning","llm","llm-evaluation","llm-inference","llm-training","mathematics","nlp"],"archived":false,"github_pushed_at":"2024-12-10T04:55:00+00:00","maintenance_label":"Dormant","url":"https://www.graphcanon.com/tools/hkust-nlp-dart-math","markdown_url":"https://www.graphcanon.com/tools/hkust-nlp-dart-math.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/hkust-nlp-dart-math","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=hkust-nlp-dart-math","description":"[NeurIPS'24] Official code for *🎯DART-Math: Difficulty-Aware Rejection Tuning for Mathematical Problem-Solving*","homepage_url":"https://hkust-nlp.github.io/dart-math/","license":"MIT","open_issues":5,"watchers":0,"ai_summary":null,"readme_excerpt":"# 🎯DART-Math\n\n\n\n\n> Difficulty-Aware Rejection Tuning for Mathematical Problem-Solving\n> \\[NeurIPS 2024\\]\n>\n> [Yuxuan Tong](https://tongyx361.github.io), Xiwen Zhang, Rui Wang,\n> Ruidong Wu, [Junxian He](https://jxhe.github.io)\n\n📝 [Paper@arXiv](https://arxiv.org/abs/2407.13690) \\| 🤗\n[Datasets&Models@HF](https://huggingface.co/collections/hkust-nlp/dart-math-665704599b35de59f8fdf6c1)\n\\| 🐱 [Code@GitHub](https://github.com/hkust-nlp/dart-math) \\| 💡\n[Slides](https://docs.google.com/presentation/d/1ZBPsM5Ww3XbQo3zAE6y-lpfsWLsK6cAbCsF6lbNHDY4/edit?usp=sharing)\n\\| 🏆 [Published@NeurIPS\n2024](https://nips.cc/virtual/2024/poster/92959)\n\n🐦\n[Thread@X(Twitter)](https://x.com/tongyx361/status/1811413243350454455)\n\\| 🐶 [中文博客@知乎](https://zhuanlan.zhihu.com/p/708371895) \\| 📊\n[Leaderboard@PapersWithCode](https://paperswithcode.com/paper/dart-math-difficulty-aware-rejection-tuning#results)\n\\| 📑\n[BibTeX](https://github.com/hkust-nlp/dart-math?tab=readme-ov-file#%EF%B8%8F-citation)\n\n> \\[!IMPORTANT\\]\n>\n> 🔥 **News!!!**\n>\n> - \\[2024/09/25\\] 🎉 *DART-Math* is accepted to [*NeurIPS\n>   2024*](https://nips.cc/virtual/2024/poster/92959)!\n> - \\[2024/07/21\\] Excited to find **our [`DART-Math-DSMath-7B`\n>   (Prop2Diff)](https://huggingface.co/hkust-nlp/dart-math-dsmath-7b-prop2diff)\n>   [comparable](https://github.com/project-numina/aimo-progress-prize/blob/main/report/numina_dataset.pdf)\n>   to the AIMO winner\n>   [NuminaMath-7B](https://huggingface.co/AI-MO/NuminaMath-7B-CoT)** on\n>   CoT, but based solely on\n>   [MATH](https://huggingface.co/datasets/hkust-nlp/dart-math-pool-math-query-info)\n>   &\n>   [GSM8K](https://huggingface.co/datasets/hkust-nlp/dart-math-pool-gsm8k-query-info)\n>   prompt set, leaving much room to improve! Besides, our [`DART`\n>   method](https://github.com/hkust-nlp/dart-math?tab=readme-ov-file#dars--difficulty-aware-rejection-sampling)\n>   is also fully [compatible with tool-integrated\n>   reasoning](https://github.com/hkust-nlp/dart-math?tab=readme-ov-file#tool-integrated-reasoning-reasoning-in-natural-language-interleaved-with-python-code).\n>   Join the discussion under this [X\n>   thread](https://x.com/tongyx361/status/1815112376649134172)!\n\n\n\n<div align=\"center\">\n\n<img src=\"https://tongyx361.github.io/assets/dart-math/main-results.png\" alt=\"Main results averaged on 2 in-domain and 4 challenging out-of-domain mathematical reasoning benchmarks.\" height=300px>\n<img src=\"https://tongyx361.github.io/assets/dart-math/main-nresp-vs-query.png\" alt=\"Number of responses v.s. query descending in difficulty in DART-Math datasets and similar-sized VRT baseline\" height=300px>\n\n</div>\n\n<div align=\"left\">\n\n<sup> Figure 1: <strong>Left:</strong> Average accuracy on 6\nmathematical benchmarks. We compare with models fine-tuned on the best,\npublic instruction tuning datasets for mathematical problem-solving:\nMetaMath <a href=\"https://openreview.net/forum?id=N8N0hgNDRt\">(Yu et\nal., 2024)</a> with 395K examples, MMIQC\n<a href=\"https://arxiv.org/abs/2401.09003\">(Liu et al., 2024a)</a> with\n2.3 million examples, as well as vanilla rejection tuning (VRT) with\n590K examples. Both <em>DART-Math (Uniform)</em> and <em>DART-Math\n(Prop2Diff)</em> use 590K training examples. <strong>Right:</strong>\nNumber of responses for each query descending by difficulty across 3\nsynthesis strategies. Queries are from the MATH training split\n<a href=\"https://datasets-benchmarks-proceedings.neurips.cc/paper/2021/hash/be83ab3ecd0db773eb2dc1b0a17836a1-Abstract-round2.html\">(Hendrycks\net al., 2021)</a>. VRT is the baseline biased towards easy queries,\nwhile <em>Uniform</em> and <em>Prop2Diff</em> are proposed in this work\nto balance and bias towards difficult queries respectively. Points are\nslightly shifted and downsampled for clarity. </sup>\n\n</div>\n\n| Dataset | Setting | \\# of Samples | [MATH](https://huggingface.co/datasets/hendrycks/competition_math) | [GSM8K](https://huggingface.co/datasets/gsm8k) | [College](https://github.com/hkust-nlp/dart-math/tree/main/d","github_created_at":"2024-05-29T09:13:03+00:00","created_at":"2026-07-11T12:02:42.12164+00:00","updated_at":"2026-07-11T12:02:56.527089+00:00","categories":[{"slug":"llm-frameworks","name":"LLM Frameworks","url":"https://www.graphcanon.com/categories/llm-frameworks","markdown_url":"https://www.graphcanon.com/categories/llm-frameworks.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/llm-frameworks"},{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"},{"slug":"inference-serving","name":"Inference & Serving","url":"https://www.graphcanon.com/categories/inference-serving","markdown_url":"https://www.graphcanon.com/categories/inference-serving.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/inference-serving"}],"tags":[{"slug":"deep-learning","name":"deep-learning"},{"slug":"llm","name":"llm"},{"slug":"nlp","name":"nlp"},{"slug":"jupyter-notebook","name":"jupyter notebook"},{"slug":"llm-inference","name":"llm-inference"},{"slug":"mathematics","name":"mathematics"},{"slug":"llm-training","name":"llm-training"},{"slug":"llm-evaluation","name":"llm-evaluation"}],"trust":{"provenance":{"is_fork":false,"github_id":807526338,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-07-11T12:02:42.719Z","maintenance":{"label":"Dormant","score":18,"methodology":"github_public_v1","releases_90d":0,"days_since_push":578,"last_release_at":null},"security_summary":{"status":"ok","scanner":"osv@v1","low_count":0,"high_count":0,"last_scan_at":"2026-07-11T12:02:45.365Z","medium_count":0,"scan_profile":"deps","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-07-11T12:02:44.982Z"},"languages":{"value":["jupyter notebook"],"source":"github.language","observed_at":"2026-07-11T12:02:44.982Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-07-11T12:02:44.982Z"}}}}