Analyzing LLMs' Knowledge Boundary Cognition Across Languages Through the Lens of Internal Representations

Xiao, Chenghao; Chan, Hou Pong; Zhang, Hao; Aljunied, Mahani; Bing, Lidong; Al Moubayed, Noura; Rong, Yu

Abstract

While understanding the knowledge boundaries of LLMs is crucial to prevent hallucination, research on the knowledge boundaries of LLMs has predominantly focused on English. In this work, we present the first study to analyze how LLMs recognize knowledge boundaries across different languages by probing their internal representations when processing known and unknown questions in multiple languages. Our empirical studies reveal three key findings: 1) LLMs' perceptions of knowledge boundaries are encoded in the middle to middle-upper layers across different languages. 2) Language differences in knowledge boundary perception follow a linear structure, which motivates our proposal of a training-free alignment method that effectively transfers knowledge boundary perception ability across languages, thereby helping reduce hallucination risk in low-resource languages; 3) Fine-tuning on bilingual question pair translation further enhances LLMs' recognition of knowledge boundaries across languages. Given the absence of standard testbeds for cross-lingual knowledge boundary analysis, we construct a multilingual evaluation suite comprising three representative types of knowledge boundary data. Our code and datasets are publicly available at https://github.com/DAMO-NLP-SG/ LLM-Multilingual-Knowledge-Boundaries.

Citation

Xiao, C., Chan, H. P., Zhang, H., Aljunied, M., Bing, L., Al Moubayed, N., & Rong, Y. (2025, July). Analyzing LLMs' Knowledge Boundary Cognition Across Languages Through the Lens of Internal Representations. Presented at Annual Meeting of the Association for Computational Linguistics (ACL), Vienna, Austria

• Presentation Conference Type: Conference Paper (published)
• Conference Name: Annual Meeting of the Association for Computational Linguistics (ACL)
• Start Date: Jul 27, 2025
• End Date: Aug 1, 2025
• Acceptance Date: Jun 2, 2025
• Deposit Date: Jun 18, 2025
• Peer Reviewed: Peer Reviewed
• Public URL: https://durham-repository.worktribe.com/output/4106988
• Publisher URL: https://aclanthology.org/events/acl-2024/