Large Language Models have Intrinsic Self-Correction Ability
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866929642996236288 |
|---|---|
| author | Liu, Dancheng Nassereldine, Amir Yang, Ziming Xu, Chenhui Hu, Yuting Li, Jiajie Kumar, Utkarsh Lee, Changjae Qin, Ruiyang Shi, Yiyu Xiong, Jinjun |
| author_facet | Liu, Dancheng Nassereldine, Amir Yang, Ziming Xu, Chenhui Hu, Yuting Li, Jiajie Kumar, Utkarsh Lee, Changjae Qin, Ruiyang Shi, Yiyu Xiong, Jinjun |
| contents | Large language models (LLMs) have attracted significant attention for their exceptional abilities in various natural language processing tasks, but they suffer from hallucinations that will cause performance degradation. One promising solution to improve the LLMs' performance is to ask LLMs to revise their answer after generation, a technique known as self-correction. Among the two types of self-correction, intrinsic self-correction is considered a promising direction because it does not utilize external knowledge. However, recent works doubt the validity of LLM's ability to conduct intrinsic self-correction. In this paper, we present a novel perspective on the intrinsic self-correction capabilities of LLMs through theoretical analyses and empirical experiments. In addition, we identify two critical factors for successful self-correction: zero temperature and fair prompts. Leveraging these factors, we demonstrate that intrinsic self-correction ability is exhibited across multiple existing LLMs. Our findings offer insights into the fundamental theories underlying the self-correction behavior of LLMs and remark on the importance of unbiased prompts and zero temperature settings in harnessing their full potential. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_15673 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Large Language Models have Intrinsic Self-Correction Ability Liu, Dancheng Nassereldine, Amir Yang, Ziming Xu, Chenhui Hu, Yuting Li, Jiajie Kumar, Utkarsh Lee, Changjae Qin, Ruiyang Shi, Yiyu Xiong, Jinjun Computation and Language Artificial Intelligence Large language models (LLMs) have attracted significant attention for their exceptional abilities in various natural language processing tasks, but they suffer from hallucinations that will cause performance degradation. One promising solution to improve the LLMs' performance is to ask LLMs to revise their answer after generation, a technique known as self-correction. Among the two types of self-correction, intrinsic self-correction is considered a promising direction because it does not utilize external knowledge. However, recent works doubt the validity of LLM's ability to conduct intrinsic self-correction. In this paper, we present a novel perspective on the intrinsic self-correction capabilities of LLMs through theoretical analyses and empirical experiments. In addition, we identify two critical factors for successful self-correction: zero temperature and fair prompts. Leveraging these factors, we demonstrate that intrinsic self-correction ability is exhibited across multiple existing LLMs. Our findings offer insights into the fundamental theories underlying the self-correction behavior of LLMs and remark on the importance of unbiased prompts and zero temperature settings in harnessing their full potential. |
| title | Large Language Models have Intrinsic Self-Correction Ability |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2406.15673 |