Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2512.20845 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914218010214400 |
|---|---|
| author | Ozer, Onat Wu, Grace Wang, Yuchen Dosti, Daniel Zhang, Honghao De La Rue, Vivi |
| author_facet | Ozer, Onat Wu, Grace Wang, Yuchen Dosti, Daniel Zhang, Honghao De La Rue, Vivi |
| contents | LLMs have shown the capacity to improve their performance on reasoning tasks through reflecting on their mistakes, and acting with these reflections in mind. However, continual reflections of the same LLM onto itself exhibit degeneration of thought, where the LLM continues to repeat the same errors again and again even with the knowledge that its wrong. To address this problem, we instead introduce multi-agent with multi-persona debators as the method to generate reflections. Through out extensive experimentation, we've found that the leads to better diversity of in the reflections generated by the llm agent. We demonstrate an accuracy of 47% EM HotPot QA (question answering) and 82.7% on HumanEval (programming), both performances surpassing reflection with a single llm. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_20845 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MAR:Multi-Agent Reflexion Improves Reasoning Abilities in LLMs Ozer, Onat Wu, Grace Wang, Yuchen Dosti, Daniel Zhang, Honghao De La Rue, Vivi Artificial Intelligence Multiagent Systems LLMs have shown the capacity to improve their performance on reasoning tasks through reflecting on their mistakes, and acting with these reflections in mind. However, continual reflections of the same LLM onto itself exhibit degeneration of thought, where the LLM continues to repeat the same errors again and again even with the knowledge that its wrong. To address this problem, we instead introduce multi-agent with multi-persona debators as the method to generate reflections. Through out extensive experimentation, we've found that the leads to better diversity of in the reflections generated by the llm agent. We demonstrate an accuracy of 47% EM HotPot QA (question answering) and 82.7% on HumanEval (programming), both performances surpassing reflection with a single llm. |
| title | MAR:Multi-Agent Reflexion Improves Reasoning Abilities in LLMs |
| topic | Artificial Intelligence Multiagent Systems |
| url | https://arxiv.org/abs/2512.20845 |