├── figures └── categories.png ├── LICENSE └── README.md /figures/categories.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ryokamoi/llm-self-correction-papers/HEAD/figures/categories.png -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "[]" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright [yyyy] [name of copyright owner] 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # LLM Self-Correction Papers 2 | 3 | This repository contains a list of papers on self-correction of large language models (LLMs). 4 | 5 | The list is maintained by [Ryo Kamoi](https://ryokamoi.github.io/). If you have any suggestions or corrections, please feel free to open an issue or a pull request (refer to [Contributing](#contributing) for details). 6 | 7 | This list is based on [our survey paper](https://arxiv.org/abs/2406.01297). If you find this list useful, please consider citing our paper: 8 | 9 | ```bibtex 10 | @article{kamoi2024self-correction, 11 | author = {Kamoi, Ryo and Zhang, Yusen and Zhang, Nan and Han, Jiawei and Zhang, Rui}, 12 | title = "{When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMs}", 13 | journal = {Transactions of the Association for Computational Linguistics}, 14 | volume = {12}, 15 | pages = {1417-1440}, 16 | year = {2024}, 17 | month = {11}, 18 | issn = {2307-387X}, 19 | doi = {10.1162/tacl_a_00713}, 20 | url = {https://doi.org/10.1162/tacl\_a\_00713}, 21 | eprint = {https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl\_a\_00713/2478635/tacl\_a\_00713.pdf}, 22 | } 23 | ``` 24 | 25 | ## Self-Correction of LLMs 26 | 27 | Self-correction of LLMs is a framework that refines responses from LLMs using LLMs during inference. Previous work has proposed various frameworks for self-correction, such as using external tools or information, or training LLMs specifically for self-correction. 28 | 29 |

30 | 31 |

32 | 33 | In this repository, we focus on inference-time self-correction, and differentiate it from training-time self-improvement of LLMs, which uses their own responses for improving themselves only during training. 34 | 35 | We also do not cover generate-and-rank (or sample-and-rank), which generate multiple responses and rank them using LLMs or other models. In contrast, self-correction refines their own responses, not only selecting the best one from multiple responses. 36 | 37 | ## Table of Contents 38 | 39 | - [Survey Papers of Self-Correction](#survey-papers-of-self-correction) 40 | - [Intrinsic Self-Correction](#intrinsic-self-correction) 41 | - [Negative Results of Intrinsic Self-Correction](#negative-results-of-intrinsic-self-correction) 42 | - [Self-Correction with External Tools](#self-correction-with-external-tools) 43 | - [with In-Context Learning (External Tools)](#with-in-context-learning-external-tools) 44 | - [with Training (External Tools)](#with-training-external-tools) 45 | - [Self-Correction with Information Retrieval](#self-correction-with-information-retrieval) 46 | - [with In-Context Learning (Information Retrieval)](#with-in-context-learning-information-retrieval) 47 | - [with Training (Information Retrieval)](#with-training-information-retrieval) 48 | - [Self-Correction with Training Designed for Self-Correction](#self-correction-with-training-designed-for-self-correction) 49 | - [Supervised Fine-tuning w/ GT feedback](#supervised-fine-tuning-w-gt-feedback) 50 | - [Supervised Fine-tuning w/o GT feedback](#supervised-fine-tuning-wo-gt-feedback) 51 | - [Reinforcement Learning](#reinforcement-learning) 52 | - [o1-like Frameworks](#o1-like-frameworks) 53 | - [Self-Correction of Vision Language Models](#self-correction-of-vision-language-models) 54 | - [Contributing](#contributing) 55 | - [License](#license) 56 | 57 | ## Survey Papers of Self-Correction 58 | 59 | * **Automatically Correcting Large Language Models: Surveying the Landscape of Diverse Automated Correction Strategies.** *Liangming Pan, Michael Saxon, Wenda Xu, Deepak Nathani, Xinyi Wang, and William Yang Wang.* TACL. 2024. [[paper](https://doi.org/10.1162/tacl_a_00660)] [[paper list](https://github.com/teacherpeterpan/self-correction-llm-papers)] 60 | * **When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMs.** *Ryo Kamoi, Yusen Zhang, Nan Zhang, Jiawei Han, and Rui Zhang.* TACL. 2024. [[paper](https://doi.org/10.1162/tacl_a_00713)] 61 | * **A Survey on LLM Inference-Time Self-Improvement.** *Xiangjue Dong, Maria Teleki, James Caverlee.* Preprint. 2024. [[paper](https://www.arxiv.org/abs/2412.14352)] 62 | 63 | ## Intrinsic Self-Correction 64 | 65 | Intrinsic self-correction is a framework that refines responses from LLMs using the same LLMs without using external feedback or training designed for self-correction. 66 | 67 | * **Constitutional AI: Harmlessness from AI Feedback.** *Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, Carol Chen, Catherine Olsson, Christopher Olah, Danny Hernandez, Dawn Drain, Deep Ganguli, Dustin Li, Eli Tran-Johnson, Ethan Perez, Jamie Kerr, Jared Mueller, Jeffrey Ladish, Joshua Landau, Kamal Ndousse, Kamile Lukosuite, Liane Lovitt, Michael Sellitto, Nelson Elhage, Nicholas Schiefer, Noemi Mercado, Nova DasSarma, Robert Lasenby, Robin Larson, Sam Ringer, Scott Johnston, Shauna Kravec, Sheer El Showk, Stanislav Fort, Tamera Lanham, Timothy Telleen-Lawton, Tom Conerly, Tom Henighan, Tristan Hume, Samuel R. Bowman, Zac Hatfield-Dodds, Ben Mann, Dario Amodei, Nicholas Joseph, Sam McCandlish, Tom Brown, and Jared Kaplan.* Preprint. 2022. [[paper](https://arxiv.org/abs/2212.08073)] 68 | * [Self-Refine] **Self-Refine: Iterative Refinement with Self-Feedback.** *Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, Shashank Gupta, Bodhisattwa Prasad Majumder, Katherine Hermann, Sean Welleck, Amir Yazdanbakhsh, and Peter Clark.* NeurIPS. 2023. [[paper](https://proceedings.neurips.cc/paper_files/paper/2023/file/91edff07232fb1b55a505a9e9f6c0ff3-Paper-Conference.pdf)] 69 | * [CoVe] **Chain-of-Verification Reduces Hallucination in Large Language Models.** *Shehzaad Dhuliawala, Mojtaba Komeili, Jing Xu, Roberta Raileanu, Xian Li, Asli Celikyilmaz, and Jason Weston.* Preprint. 2023. [[paper](https://arxiv.org/abs/2309.11495)] 70 | * [RCI] **Language Models can Solve Computer Tasks.** *Geunwoo Kim, Pierre Baldi, and Stephen McAleer.* Preprint. 2023. [[paper](https://arxiv.org/abs/2303.17491)] 71 | * [Reflexion] **Reflexion: language agents with verbal reinforcement learning.** *Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao.* NeurIPS. 2023. [[paper](https://proceedings.neurips.cc/paper_files/paper/2023/file/1b44b878bb782e6954cd888628510e90-Paper-Conference.pdf)] 72 | 73 | 74 | ### Negative Results of Intrinsic Self-Correction 75 | 76 | It has been reported that intrinsic self-correction does not work well in many tasks. 77 | 78 | * **Large Language Models Cannot Self-Correct Reasoning Yet.** *Jie Huang, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng, Adams Wei Yu, Xinying Song, and Denny Zhou.* ICLR. 2024. [[paper](https://openreview.net/forum?id=IkmD3fKBPQ)] 79 | * [CRITIC] **CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing.** *Zhibin Gou, Zhihong Shao, Yeyun Gong, Yelong Shen, Yujiu Yang, Nan Duan, and Weizhu Chen.* ICLR. 2024. [[paper](https://openreview.net/forum?id=Sx038qxjek)] 80 | * **When is Tree Search Useful for LLM Planning? It Depends on the Discriminator.** *Ziru Chen, Michael White, Raymond Mooney, Ali Payani, Yu Su, and Huan Sun.* ACL. 2024. [[paper](https://arxiv.org/abs/2402.10890)] [[code](https://github.com/OSU-NLP-Group/llm-planning-eval)] 81 | 82 | ## Self-Correction with External Tools 83 | 84 | Previous work has proposed self-correction frameworks that use external tools, such as code executors for code generation tasks and proof assistants for theorem proving tasks. 85 | 86 | ### with In-Context Learning (External Tools) 87 | 88 | * [Reflexion] **Reflexion: language agents with verbal reinforcement learning.** *Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao.* NeurIPS. 2023. [[paper](https://proceedings.neurips.cc/paper_files/paper/2023/file/1b44b878bb782e6954cd888628510e90-Paper-Conference.pdf)] 89 | * [SelfEvolve] **SelfEvolve: A Code Evolution Framework via Large Language Models.** *Shuyang Jiang, Yuhao Wang, and Yu Wang.* Preprint. 2023. [[paper](https://arxiv.org/abs/2306.02907)] 90 | * [Logic-LM] **Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoning.** *Liangming Pan, Alon Albalak, Xinyi Wang, and William Wang.* Findings of EMNLP. 2023. [[paper](https://aclanthology.org/2023.findings-emnlp.248)] 91 | * [Self-Debug] **Teaching Large Language Models to Self-Debug.** *Xinyun Chen, Maxwell Lin, Nathanael Schärli, and Denny Zhou.* ICLR. 2024. [[paper](https://openreview.net/forum?id=KuPixIqPiq)] 92 | * [CRITIC] **CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing.** *Zhibin Gou, Zhihong Shao, Yeyun Gong, Yelong Shen, Yujiu Yang, Nan Duan, and Weizhu Chen.* ICLR. 2024. [[paper](https://openreview.net/forum?id=Sx038qxjek)] 93 | 94 | ### with Training (External Tools) 95 | 96 | * [CodeRL] **CodeRL: Mastering Code Generation through Pretrained Models and Deep Reinforcement Learning.** *Hung Le, Yue Wang, Akhilesh Deepak Gotmare, Silvio Savarese, and Steven Chu Hong Hoi.* NeurIPS. 2022. [[paper](https://proceedings.neurips.cc/paper_files/paper/2022/file/8636419dea1aa9fbd25fc4248e702da4-Paper-Conference.pdf)] 97 | * [Self-Edit] **Self-Edit: Fault-Aware Code Editor for Code Generation.** *Kechi Zhang, Zhuo Li, Jia Li, Ge Li, and Zhi Jin.* ACL. 2023. [[paper](https://aclanthology.org/2023.acl-long.45)] 98 | * [Baldur] **Baldur: Whole-Proof Generation and Repair with Large Language Models.** *Emily First, Markus Rabe, Talia Ringer, and Yuriy Brun.* ESEC/FSE. 2023. [[paper](https://doi.org/10.1145/3611643.3616243)] 99 | 100 | ## Self-Correction with Information Retrieval 101 | 102 | Previous work has proposed self-correction frameworks that use information retrieval during inference. 103 | 104 | ### with In-Context Learning (Information Retrieval) 105 | 106 | * [RARR] **RARR: Researching and Revising What Language Models Say, Using Language Models.** *Luyu Gao, Zhuyun Dai, Panupong Pasupat, Anthony Chen, Arun Tejasvi Chaganty, Yicheng Fan, Vincent Zhao, Ni Lao, Hongrae Lee, Da-Cheng Juan, and Kelvin Guu.* ACL. 2023. [[paper](https://aclanthology.org/2023.acl-long.910)] 107 | * [Verify-and-Edit] **Verify-and-Edit: A Knowledge-Enhanced Chain-of-Thought Framework.** *Ruochen Zhao, Xingxuan Li, Shafiq Joty, Chengwei Qin, and Lidong Bing.* ACL. 2023. [[paper](https://aclanthology.org/2023.acl-long.320)] 108 | * **A Stitch in Time Saves Nine: Detecting and Mitigating Hallucinations of LLMs by Validating Low-Confidence Generation.** *Neeraj Varshney, Wenlin Yao, Hongming Zhang, Jianshu Chen, and Dong Yu.* Preprint. 2023. [[paper](https://arxiv.org/abs/2307.03987)] 109 | * **Improving Language Models via Plug-and-Play Retrieval Feedback.** *Wenhao Yu, Zhihan Zhang, Zhenwen Liang, Meng Jiang, and Ashish Sabharwal.* Preprint. 2023. [[paper](https://arxiv.org/abs/2305.14002)] 110 | * [CRITIC] **CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing.** *Zhibin Gou, Zhihong Shao, Yeyun Gong, Yelong Shen, Yujiu Yang, Nan Duan, and Weizhu Chen.* ICLR. 2024. [[paper](https://openreview.net/forum?id=Sx038qxjek)] 111 | * [FLARE] **Active Retrieval Augmented Generation.** *Zhengbao Jiang, Frank Xu, Luyu Gao, Zhiqing Sun, Qian Liu, Jane Dwivedi-Yu, Yiming Yang, Jamie Callan, and Graham Neubig.* EMNLP. 2023. [[paper](https://aclanthology.org/2023.emnlp-main.495)] 112 | 113 | ### with Training (Information Retrieval) 114 | 115 | * 116 | 117 | ## Self-Correction with Training Designed for Self-Correction 118 | 119 | This section includes self-correction frameworks that train LLMs specifically for self-correction, but do not use external tools or information retrieval during inference. 120 | 121 | ### Supervised Fine-tuning (w/ GT feedback) 122 | 123 | It has been reported that LLMs often can self-correct their own mistakes when fine-tuned on ground-truth feedback (e.g., human annotated or generated by stronger models). 124 | 125 | * **Self-critiquing models for assisting human evaluators.** *William Saunders, Catherine Yeh, Jeff Wu, Steven Bills, Long Ouyang, Jonathan Ward, and Jan Leike.* Preprint. 2022. [[paper](https://arxiv.org/abs/2206.05802)] 126 | * [Re3] **Re3: Generating Longer Stories With Recursive Reprompting and Revision.** *Kevin Yang, Yuandong Tian, Nanyun Peng, and Dan Klein.* EMNLP. 2022. [[paper](https://aclanthology.org/2022.emnlp-main.296)] 127 | * [SelFee] **SelFee: Iterative Self-Revising LLM Empowered by Self-Feedback Generation.** *Seonghyeon Ye, Yongrae Jo, Doyoung Kim, Sungdong Kim, Hyeonbin Hwang, and Minjoon Seo.* Blog post. 2023. [[blog](https://kaistai.github.io/SelFee/)] 128 | * [Volcano] **Volcano: Mitigating Multimodal Hallucination through Self-Feedback Guided Revision.** *Seongyun Lee, Sue Park, Yongrae Jo, and Minjoon Seo.* NAACL. 2024. [[paper](https://aclanthology.org/2024.naacl-long.23)] 129 | * [REFINER] **REFINER: Reasoning Feedback on Intermediate Representations.** *Debjit Paul, Mete Ismayilzada, Maxime Peyrard, Beatriz Borges, Antoine Bosselut, Robert West, and Boi Faltings.* EACL. 2024. [[paper](https://aclanthology.org/2024.eacl-long.67)] 130 | * [GLoRe] **GLoRe: When, Where, and How to Improve LLM Reasoning via Global and Local Refinements.** *Alex Havrilla, Sharath Raparthy, Christoforus Nalmpantis, Jane Dwivedi-Yu, Maksym Zhuravinskyi, Eric Hambro, and Roberta Raileanu.* Preprint. 2024. [[paper](https://arxiv.org/abs/2402.10963)]. 131 | 132 | ### Supervised Fine-tuning (w/o GT feedback) 133 | 134 | There are studies that attempt to improve self-correction without using ground-truth feedback because human annotation for self-correction is costly. 135 | 136 | * [Self-corrective learning] **Generating Sequences by Learning to Self-Correct.** *Sean Welleck, Ximing Lu, Peter West, Faeze Brahman, Tianxiao Shen, Daniel Khashabi, and Yejin Choi.* ICLR. 2023. [[paper](https://openreview.net/forum?id=hH36JeQZDaO)] 137 | * **Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters.** *Charlie Snell, Jaehoon Lee, Kelvin Xu, and Aviral Kumar.* Preprint. 2024. [[paper](https://arxiv.org/abs/2408.03314)]. 138 | * [Meta-Reflection] **Meta-Reflection: A Feedback-Free Reflection Learning Framework.** *Yaoke Wang, Yun Zhu, Xintong Bao, Wenqiao Zhang, Suyang Dai, Kehan Chen, Wenqiang Li, Gang Huang, Siliang Tang, Yueting Zhuang.* Preprint. 2024. [[paper](https://arxiv.org/abs/2412.13781)] 139 | 140 | ### Reinforcement Learning 141 | 142 | Reinforcement learning is another approach to improve self-correction without using human annotated feedback. 143 | 144 | * [RL4F] **RL4F: Generating Natural Language Feedback with Reinforcement Learning for Repairing Model Outputs.** *Afra Feyza Akyurek, Ekin Akyurek, Ashwin Kalyan, Peter Clark, Derry Tanti Wijaya, and Niket Tandon.* ACL. 2023. [[paper](https://aclanthology.org/2023.acl-long.427)] 145 | * [RISE] **Recursive Introspection: Teaching Language Model Agents How to Self-Improve.** *Yuxiao Qu, Tianjun Zhang, Naman Garg, and Aviral Kumar.* Preprint. 2024. [[paper](https://arxiv.org/abs/2407.18219)] 146 | * [SCoRe] **Training Language Models to Self-Correct via Reinforcement Learning.** *Aviral Kumar, Vincent Zhuang, Rishabh Agarwal, Yi Su, John D Co-Reyes, Avi Singh, Kate Baumli, Shariq Iqbal, Colton Bishop, Rebecca Roelofs, Lei M Zhang, Kay McKinney, Disha Shrivastava, Cosmin Paduraru, George Tucker, Doina Precup, Feryal Behbahani, and Aleksandra Faust.* Preprint. 2024. [[paper](https://arxiv.org/abs/2409.12917)] 147 | 148 | ### o1-like Frameworks 149 | 150 | OpenAI o1 is a framework focusing on improving the reasoning capabilities of LLMs trained with **reinforcement learning** to explore multiple reasoning processes and correct their own mistakes during inference. After the release of OpenAI o1, several papers or projects have proposed frameworks similar to OpenAI o1. 151 | 152 | * [OpenAI o1] **Learning to Reason with LLMs.** *OpenAI.* Blog post. 2024. [[blog](https://openai.com/index/learning-to-reason-with-llms/)] [[system card](https://cdn.openai.com/o1-system-card-20241205.pdf)] 153 | * **Evaluation of OpenAI o1: Opportunities and Challenges of AGI.** *Tianyang Zhong, Zhengliang Liu, Yi Pan, Yutong Zhang, Yifan Zhou, Shizhe Liang, Zihao Wu, Yanjun Lyu, Peng Shu, Xiaowei Yu, Chao Cao, Hanqi Jiang, Hanxu Chen, Yiwei Li, Junhao Chen, Huawen Hu, Yihen Liu, Huaqin Zhao, Shaochen Xu, Haixing Dai, Lin Zhao, Ruidong Zhang, Wei Zhao, Zhenyuan Yang, Jingyuan Chen, Peilong Wang, Wei Ruan, Hui Wang, Huan Zhao, Jing Zhang, Yiming Ren, Shihuan Qin, Tong Chen, Jiaxi Li, Arif Hassan Zidan, Afrar Jahin, Minheng Chen, Sichen Xia, Jason Holmes, Yan Zhuang, Jiaqi Wang, Bochen Xu, Weiran Xia, Jichao Yu, Kaibo Tang, Yaxuan Yang, Bolun Sun, Tao Yang, Guoyu Lu, Xianqiao Wang, Lilong Chai, He Li, Jin Lu, Lichao Sun, Xin Zhang, Bao Ge, Xintao Hu, Lian Zhang, Hua Zhou, Lu Zhang, Shu Zhang, Ninghao Liu, Bei Jiang, Linglong Kong, Zhen Xiang, Yudan Ren, Jun Liu, Xi Jiang, Yu Bao, Wei Zhang, Xiang Li, Gang Li, Wei Liu, Dinggang Shen, Andrea Sikora, Xiaoming Zhai, Dajiang Zhu, and Tianming Liu.* Preprint. 2024. [[paper](https://arxiv.org/abs/2409.18486)]. 154 | * [Skywork-o1] **Skywork-o1 Open Series.** *Skywork-o1 Team.* Models. 2024. [[models](https://huggingface.co/Skywork)]. 155 | * [LLaVA-CoT] **LLaVA-CoT: Let Vision Language Models Reason Step-by-Step.** *Guowei Xu, Peng Jin, Hao Li, Yibing Song, Lichao Sun, and Li Yuan.* Preprint. 2024. [[paper](https://arxiv.org/abs/2411.10440)]. 156 | * [Marco-o1] **Marco-o1: Towards Open Reasoning Models for Open-Ended Solutions.** *Yu Zhao, Huifeng Yin, Bo Zeng, Hao Wang, Tianqi Shi, Chenyang Lyu, Longyue Wang, Weihua Luo, and Kaifu Zhang.* Preprint. 2024. [[paper](https://arxiv.org/abs/2411.14405)] 157 | * [QwQ] **QwQ: Reflect Deeply on the Boundaries of the Unknown.** *Qwen Team.* Blog post. 2024. [[blog](https://qwenlm.github.io/blog/qwq-32b-preview/)] 158 | 159 | For more papers related to OpenAI o1, please also refer to the following repositories. 160 | 161 | * [https://github.com/srush/awesome-o1](https://github.com/srush/awesome-o1) 162 | * [https://github.com/hijkzzz/Awesome-LLM-Strawberry](https://github.com/hijkzzz/Awesome-LLM-Strawberry) 163 | 164 | ## Self-Correction of Vision Language Models 165 | 166 | * [Volcano] **Volcano: Mitigating Multimodal Hallucination through Self-Feedback Guided Revision.** *Seongyun Lee, Sue Park, Yongrae Jo, and Minjoon Seo.* NAACL. 2024. [[paper](https://aclanthology.org/2024.naacl-long.23)] 167 | * [VISCO] **VISCO: Benchmarking Fine-Grained Critique and Correction Towards Self-Improvement in Visual Reasoning.** *Xueqing Wu, Yuheng Ding, Bingxuan Li, Pan Lu, Da Yin, Kai-Wei Chang, and Nanyun Peng.* Preprint. 2024. [[paper](https://arxiv.org/abs/2412.02172)]. 168 | 169 | ## Contributing 170 | 171 | We welcome contributions! If you’d like to add a new paper to this list, please submit a pull request. **Ensure that your commit and PR have descriptive and unique titles** rather than generic ones like "Updated README.md." 172 | 173 | Kindly use the following format for your entry: 174 | 175 | ```md 176 | * [Short name (if exists)] **Paper Title.** *Author1, Author2, ... , and Last Author.* Conference/Journal/Preprint/Blog post. year. [[paper](url)] [[code, follow up paper, etc.](url)] 177 | ``` 178 | 179 | If you have any questions or suggestions, please feel free to open an issue or reach out to [Ryo Kamoi](https://ryokamoi.github.io/) (ryokamoi@psu.edu). 180 | 181 | ## License 182 | 183 | Please refer to [LICENSE](LICENSE). 184 | --------------------------------------------------------------------------------