Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2406.14434 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929736922431488 |
|---|---|
| author | Liu, Weihao Wu, Ning Ding, Wenbiao Liang, Shining Gong, Ming Zhang, Dongmei |
| author_facet | Liu, Weihao Wu, Ning Ding, Wenbiao Liang, Shining Gong, Ming Zhang, Dongmei |
| contents | Truthfulness stands out as an essential challenge for Large Language Models (LLMs). Although many works have developed various ways for truthfulness enhancement, they seldom focus on truthfulness in multilingual scenarios. Meanwhile, contemporary multilingual aligning technologies struggle to balance numerous languages and often exhibit serious truthfulness gaps across different languages, especially those that differ greatly from English. In our work, we extend truthfulness evaluation to multilingual contexts and propose a practical method for cross-lingual truthfulness transfer called Fact-aware Multilingual Selective Synergy (FaMSS). FaMSS is able to select an optimal subset of all tested languages by language bias and transfer contributions, and then employ translation instruction tuning for cross-lingual truthfulness transfer. Experimental results demonstrate that our approach can effectively reduce the multilingual representation disparity and boost cross-lingual truthfulness transfer of LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_14434 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Selected Languages are All You Need for Cross-lingual Truthfulness Transfer Liu, Weihao Wu, Ning Ding, Wenbiao Liang, Shining Gong, Ming Zhang, Dongmei Computation and Language Truthfulness stands out as an essential challenge for Large Language Models (LLMs). Although many works have developed various ways for truthfulness enhancement, they seldom focus on truthfulness in multilingual scenarios. Meanwhile, contemporary multilingual aligning technologies struggle to balance numerous languages and often exhibit serious truthfulness gaps across different languages, especially those that differ greatly from English. In our work, we extend truthfulness evaluation to multilingual contexts and propose a practical method for cross-lingual truthfulness transfer called Fact-aware Multilingual Selective Synergy (FaMSS). FaMSS is able to select an optimal subset of all tested languages by language bias and transfer contributions, and then employ translation instruction tuning for cross-lingual truthfulness transfer. Experimental results demonstrate that our approach can effectively reduce the multilingual representation disparity and boost cross-lingual truthfulness transfer of LLMs. |
| title | Selected Languages are All You Need for Cross-lingual Truthfulness Transfer |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2406.14434 |