IRCAN: Mitigating Knowledge Conflicts in LLM Generation via Identifying and Reweighting Context-Aware Neurons
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_ | 1866917837021380608 |
|---|---|
| author | Shi, Dan Jin, Renren Shen, Tianhao Dong, Weilong Wu, Xinwei Xiong, Deyi |
| author_facet | Shi, Dan Jin, Renren Shen, Tianhao Dong, Weilong Wu, Xinwei Xiong, Deyi |
| contents | It is widely acknowledged that large language models (LLMs) encode a vast reservoir of knowledge after being trained on mass data. Recent studies disclose knowledge conflicts in LLM generation, wherein outdated or incorrect parametric knowledge (i.e., encoded knowledge) contradicts new knowledge provided in the context. To mitigate such knowledge conflicts, we propose a novel framework, IRCAN (Identifying and Reweighting Context-Aware Neurons) to capitalize on neurons that are crucial in processing contextual cues. Specifically, IRCAN first identifies neurons that significantly contribute to context processing, utilizing a context-aware attribution score derived from integrated gradients. Subsequently, the identified context-aware neurons are strengthened via reweighting. In doing so, we steer LLMs to generate context-sensitive outputs with respect to the new knowledge provided in the context. Extensive experiments conducted across a variety of models and tasks demonstrate that IRCAN not only achieves remarkable improvements in handling knowledge conflicts but also offers a scalable, plug-and-play solution that can be integrated seamlessly with existing models. Our codes are released at https://github.com/danshi777/IRCAN. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_18406 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | IRCAN: Mitigating Knowledge Conflicts in LLM Generation via Identifying and Reweighting Context-Aware Neurons Shi, Dan Jin, Renren Shen, Tianhao Dong, Weilong Wu, Xinwei Xiong, Deyi Computation and Language Artificial Intelligence It is widely acknowledged that large language models (LLMs) encode a vast reservoir of knowledge after being trained on mass data. Recent studies disclose knowledge conflicts in LLM generation, wherein outdated or incorrect parametric knowledge (i.e., encoded knowledge) contradicts new knowledge provided in the context. To mitigate such knowledge conflicts, we propose a novel framework, IRCAN (Identifying and Reweighting Context-Aware Neurons) to capitalize on neurons that are crucial in processing contextual cues. Specifically, IRCAN first identifies neurons that significantly contribute to context processing, utilizing a context-aware attribution score derived from integrated gradients. Subsequently, the identified context-aware neurons are strengthened via reweighting. In doing so, we steer LLMs to generate context-sensitive outputs with respect to the new knowledge provided in the context. Extensive experiments conducted across a variety of models and tasks demonstrate that IRCAN not only achieves remarkable improvements in handling knowledge conflicts but also offers a scalable, plug-and-play solution that can be integrated seamlessly with existing models. Our codes are released at https://github.com/danshi777/IRCAN. |
| title | IRCAN: Mitigating Knowledge Conflicts in LLM Generation via Identifying and Reweighting Context-Aware Neurons |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2406.18406 |