Lost in Multilinguality: Dissecting Cross-lingual Factual Inconsistency in Transformer Language Models
This addresses the issue of inconsistent factual outputs across languages for users of multilingual AI systems, but it is incremental as it builds on known inconsistency problems with a new mitigation strategy.
The paper tackled the problem of cross-lingual factual inconsistency in multilingual language models, finding that failures during language transitions in final layers cause incorrect predictions, and proposed a linear shortcut method that improved prediction accuracy and consistency.
Multilingual language models (MLMs) store factual knowledge across languages but often struggle to provide consistent responses to semantically equivalent prompts in different languages. While previous studies point out this cross-lingual inconsistency issue, the underlying causes remain unexplored. In this work, we use mechanistic interpretability methods to investigate cross-lingual inconsistencies in MLMs. We find that MLMs encode knowledge in a language-independent concept space through most layers, and only transition to language-specific spaces in the final layers. Failures during the language transition often result in incorrect predictions in the target language, even when the answers are correct in other languages. To mitigate this inconsistency issue, we propose a linear shortcut method that bypasses computations in the final layers, enhancing both prediction accuracy and cross-lingual consistency. Our findings shed light on the internal mechanisms of MLMs and provide a lightweight, effective strategy for producing more consistent factual outputs.