{"id":18887129,"url":"https://github.com/thunlp-mt/panda","last_synced_at":"2026-02-24T03:30:17.851Z","repository":{"id":223434991,"uuid":"759723281","full_name":"THUNLP-MT/PANDA","owner":"THUNLP-MT","description":null,"archived":false,"fork":false,"pushed_at":"2024-06-18T02:50:36.000Z","size":1689,"stargazers_count":1,"open_issues_count":0,"forks_count":1,"subscribers_count":2,"default_branch":"main","last_synced_at":"2024-12-31T05:18:58.728Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","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/THUNLP-MT.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}},"created_at":"2024-02-19T08:05:01.000Z","updated_at":"2024-07-22T10:38:09.000Z","dependencies_parsed_at":"2024-11-08T07:44:38.961Z","dependency_job_id":null,"html_url":"https://github.com/THUNLP-MT/PANDA","commit_stats":null,"previous_names":["xxmlala/panda"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/THUNLP-MT%2FPANDA","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/THUNLP-MT%2FPANDA/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/THUNLP-MT%2FPANDA/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/THUNLP-MT%2FPANDA/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/THUNLP-MT","download_url":"https://codeload.github.com/THUNLP-MT/PANDA/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":239859563,"owners_count":19708863,"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","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":"2024-11-08T07:34:32.066Z","updated_at":"2026-02-24T03:30:17.799Z","avatar_url":"https://github.com/THUNLP-MT.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# PANDA\nThis repo contains the codes for our work \n[🐼PANDA: Preference Adaptation for Enhancing\nDomain-Specific Abilities of LLMs](https://arxiv.org/abs/2402.12835).\n\nThe required package can be installed by running the following command.\n```sh\npip install -r requirements.txt\n```\n\n## ScienceWorld\nFirstly, switch to the ScienceWorld workspace:\n```sh\ncd ScienceWorld\n```\n\n0. To directly run experiments with PANDA (You can also run PANDA from scratch by starting from step1).\n```sh\n./run_eval_react_panda.sh\n./run_eval_reflexion_panda.sh\n./run_eval_saycan_panda.sh\n```\n1. Step1: Gather expert trials to construct preferences data.\n```sh\n./gather_trials.sh\n```\n2. Step2: PANDA-Learning from the expert preferences.\n```sh\n./panda_learning.sh\n```\n3. Step3: Test with PANDA-Insight:\n```sh\n./run_eval_react_panda.sh\n./run_eval_reflexion_panda.sh\n./run_eval_saycan_panda.sh\n```\n## TweetEval\nFirstly, switch to the ScienceWorld workspace:\n```sh\ncd TweetEval\n```\n0. Step0: Download datasets file from [cardifnlp/tweeteval](https://github.com/cardiffnlp/tweeteval) and put it in the `dataset` folder and the expert models from [cardifnlp/models](https://huggingface.co/cardiffnlp) and put it in the `models` folder.\n\n1. Step1: Gather expert trials to construct preferences data.\n```sh\n./gather_trials.sh\n```\n2. Step2: PANDA-Learning from the expert preferences.\n```sh\n./panda_learning.sh\n```\n3. Step3: Test with PANDA-Insight:\n```sh\n./eval_gpt.sh\n./eval_gpt_cot.sh\n```\n## Acknowledgement\nOur codes for scienceworld are adapted from [yuchenlin/SwiftSage](https://github.com/yuchenlin/SwiftSage). Thanks for their kind open-sourced code.\n\n## Citation \nIf you find our project helpful to your research, please consider citing:\n```bib\n@inproceedings{liu2024panda,\n  title={PANDA: Preference Adaptation for Enhancing Domain-Specific Abilities of LLMs},\n  author={Liu, An and Yang, Zonghan and Zhang, Zhenhe and Hu, Qingyuan and Li, Peng and Yan, Ming and Zhang, Ji and Huang, Fei and Liu, Yang},\n  booktitle={Findings of the Association for Computational Linguistics: ACL 2024},\n  year={2024}\n}\n```","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fthunlp-mt%2Fpanda","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fthunlp-mt%2Fpanda","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fthunlp-mt%2Fpanda/lists"}