{"id":31709967,"url":"https://github.com/wnjxyk/rpc","last_synced_at":"2026-02-16T00:32:42.684Z","repository":{"id":317332212,"uuid":"1066954991","full_name":"WNJXYK/RPC","owner":"WNJXYK","description":"Official Repository for NeurIPS 2025 Paper: \"A Theoretical Study on Bridging Internal Probability and Self-Consistency for LLM Reasoning\"","archived":false,"fork":false,"pushed_at":"2025-10-07T08:52:43.000Z","size":2120,"stargazers_count":0,"open_issues_count":1,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-10-07T10:29:41.014Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/WNJXYK.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-09-30T07:26:05.000Z","updated_at":"2025-10-04T18:29:28.000Z","dependencies_parsed_at":"2025-09-30T09:32:01.118Z","dependency_job_id":null,"html_url":"https://github.com/WNJXYK/RPC","commit_stats":null,"previous_names":["wnjxyk/rpc"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/WNJXYK/RPC","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/WNJXYK%2FRPC","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/WNJXYK%2FRPC/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/WNJXYK%2FRPC/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/WNJXYK%2FRPC/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/WNJXYK","download_url":"https://codeload.github.com/WNJXYK/RPC/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/WNJXYK%2FRPC/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":279000747,"owners_count":26082879,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-10-08T02:00:06.501Z","response_time":56,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-10-09T00:09:05.036Z","updated_at":"2025-10-09T00:09:55.029Z","avatar_url":"https://github.com/WNJXYK.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"\n# [NeurIPS 2025] A Theoretical Study on Bridging Internal Probability and Self-Consistency for LLM Reasoning\n\nOfficial Repository for NeurIPS 2025 Paper: \"A Theoretical Study on Bridging Internal Probability and Self-Consistency for LLM Reasoning\"\n\n\u003cdiv align=\"center\"\u003e\n\u003ca href=\"https://arxiv.org/pdf/2502.00511\"\u003e📄 [Paper]\u003c/a\u003e\n•\n\u003ca href=\"https://wnjxyk.github.io/RPC\"\u003e🌐 [Project]\u003c/a\u003e\n•\n\u003ca href=\"https://huggingface.co/collections/WNJXYK/mathematical-llm-reasoning-paths-68e4c4e32e3ad7fa0fcad77a\"\u003e🤗 [Data Collection]\u003c/a\u003e\n•\n\u003ca href=\"https://huggingface.co/spaces/WNJXYK/RPC\"\u003e💻 [Demo]\u003c/a\u003e\n\n\u003c/div\u003e\n\n\n## 🛠️ 1. Environment Setup\n\nWe provide two ways to create the Python environment for this repository. Please choose one of the following methods:\n\n### 1.1. Using Python virtual environment:\n\n```bash\npython -m venv rpc\nsource rpc/bin/activate\npip install -r requirements.txt \n```\n\n### 1.2. Using Conda environment:\n\n```bash\nconda create -n rpc python=3.9\nconda activate rpc\npip install -r requirements.txt\n```\n\n## 🚀 2. Reproducing Experiments\n\n### 2.1. Single Experiment\n\nRun evaluation with specific parameters:\n\n```bash\npython main.py --dataset MathOdyssey --model InternLM2-Math-Plus-7B --method RPC --K 128\n```\n\n**Parameters:**\n- `--dataset`: Choose from `MATH`, `MathOdyssey`, `AIME`, `OlympiadBench`\n- `--model`: Choose from `Deepseek-Math-RL-7B`, `InternLM2-Math-Plus-1.8B`, `InternLM2-Math-Plus-7B`\n- `--method`: Choose from `PPL` (Perplexity), `SC` (Self-Consistency), `RPC` (our method)\n- `--K`: Number of reasoning paths to sample (`128` for `MathOdyssey`, `AIME`, `OlympiadBench`, and `64` for `MATH`)\n\n### 2.2. Batch Experiments\n\nRun comprehensive evaluation across multiple settings:\n\n```bash\nbash all_exps.sh\n```\n\nThis will evaluate all method-dataset-model combinations and save results to `results.txt`.\n\n### 2.3. Hints\n\n1. If you cannot download data from Hugging Face directly, please use [Hugging Mirror](https://hf-mirror.com/) instead.\n2. It may take some time to generate the cache for checking answer equality when running each dataset for the first time.\n\n## 📚 3. BibTex\n\n```bibtex\n@inproceedings{zhou24theoretical,\n      author    = {Zhou, Zhi and Tan, Yuhao and Li, Zenan and Yao, Yuan and Guo, Lan-Zhe and Li, Yu-Feng and Ma, Xiaoxing},\n      title     = {A Theorecial Study on Bridging Internal Probability and Self-Consistency for LLM Reasoning},\n      booktitle = {Advances in Neural Information Processing Systems},\n      year      = {2025},\n    }\n```","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fwnjxyk%2Frpc","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fwnjxyk%2Frpc","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fwnjxyk%2Frpc/lists"}