Investigating Context-Faithfulness in Large Language Models: The Roles of Memory Strength and Evidence Style
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911048827666432 |
|---|---|
| author | Li, Yuepei Zhou, Kang Qiao, Qiao Nguyen, Bach Wang, Qing Li, Qi |
| author_facet | Li, Yuepei Zhou, Kang Qiao, Qiao Nguyen, Bach Wang, Qing Li, Qi |
| contents | Retrieval-augmented generation (RAG) improves Large Language Models (LLMs) by incorporating external information into the response generation process. However, how context-faithful LLMs are and what factors influence LLMs' context faithfulness remain largely unexplored. In this study, we investigate the impact of memory strength and evidence presentation on LLMs' receptiveness to external evidence. We quantify the memory strength of LLMs by measuring the divergence in LLMs' responses to different paraphrases of the same question, which is not considered by previous works. We also generate evidence in various styles to examine LLMs' behavior. Our results show that for questions with high memory strength, LLMs are more likely to rely on internal memory. Furthermore, presenting paraphrased evidence significantly increases LLMs' receptiveness compared to simple repetition or adding details. These findings provide key insights for improving retrieval-augmented generation and context-aware LLMs. Our code is available at https://github.com/liyp0095/ContextFaithful. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_10955 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Investigating Context-Faithfulness in Large Language Models: The Roles of Memory Strength and Evidence Style Li, Yuepei Zhou, Kang Qiao, Qiao Nguyen, Bach Wang, Qing Li, Qi Computation and Language Artificial Intelligence Retrieval-augmented generation (RAG) improves Large Language Models (LLMs) by incorporating external information into the response generation process. However, how context-faithful LLMs are and what factors influence LLMs' context faithfulness remain largely unexplored. In this study, we investigate the impact of memory strength and evidence presentation on LLMs' receptiveness to external evidence. We quantify the memory strength of LLMs by measuring the divergence in LLMs' responses to different paraphrases of the same question, which is not considered by previous works. We also generate evidence in various styles to examine LLMs' behavior. Our results show that for questions with high memory strength, LLMs are more likely to rely on internal memory. Furthermore, presenting paraphrased evidence significantly increases LLMs' receptiveness compared to simple repetition or adding details. These findings provide key insights for improving retrieval-augmented generation and context-aware LLMs. Our code is available at https://github.com/liyp0095/ContextFaithful. |
| title | Investigating Context-Faithfulness in Large Language Models: The Roles of Memory Strength and Evidence Style |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2409.10955 |