使用音素提示：增强 LLMs 对非拉丁字母语言的多样性支持
Prompting with Phonemes: Enhancing LLMs' Multilinguality for Non-Latin Script Languages
AI总结：
针对多语言 LLMs 在非拉丁字母语言上表现不佳的问题，提出利用音素转录作为互补信号诱导脚本不变表示，并设计 Mixed-ICL 检索策略，显著提升了拉丁与非拉丁字母语言的性能。
摘要：
尽管多语言 LLMs 在各项基准测试中取得了显著性能，但我们发现它们在当代 LLM 家族的非拉丁字母语言上依然表现不佳。这种差异源于 LLMs 使用正字法脚本进行预训练，而这些脚本以拉丁字符为主，掩盖了它们与非拉丁字母共享的音系学特征。我们提出利用音素转录作为互补信号，以诱导脚本不变表示。我们的研究表明，整合音素信号能提升非拉丁和拉丁字母语言的性能，尤其在缩小两者间的性能差距上影响显著。通过详细实验，我们证明音素和正字法脚本会为上下文学习（ICL）检索到不同示例。这促使我们提出 Mixed-ICL 检索策略，对两者进一步聚合后，相比随机 ICL 检索，我们在拉丁字母语言（高达 12.6%）和非拉丁字母语言（高达 15.1%）上均取得了显著性能提升。
英文摘要：
Although multilingual LLMs have achieved remarkable performance across benchmarks, we find they continue to underperform on non-Latin script languages across contemporary LLM families. This discrepancy arises from the fact that LLMs are pretrained with orthographic scripts, which are dominated by Latin characters that obscure their shared phonology with non-Latin scripts. We propose leveraging phonemic transcriptions as complementary signals to induce script-invariant representations. Our study demonstrates that integrating phonemic signals improves performance across both non-Latin and Latin script languages, with a particularly significant impact on closing the performance gap between the two. Through detailed experiments, we show that phonemic and orthographic scripts retrieve distinct examples for in-context learning (ICL). This motivates our proposed Mixed-ICL retrieval strategy, where further aggregation from both leads to our significant performance improvements for both Latin script languages (up to 12.6%) and non-Latin script languages (up to 15.1%) compared to randomized ICL retrieval.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
