能否映射到英语？跨语言对齐在大语言模型多语言性能中的作用
Can you map it to English? The Role of Cross-Lingual Alignment in Multilingual Performance of LLMs
AI总结：
本研究探讨了LLM中跨语言对齐对多语言NLU任务性能的影响，通过引入DALI指数验证了表示对齐在正确性中的关键作用。
摘要：
大型语言模型（LLMs）能够用多种语言回答提示，尽管它们主要是在英语数据上训练的；然而，驱动这种泛化的能力机制仍然理解不足。本研究探讨：LLM如何通过将非英语输入的表示对齐到英语来影响其在自然语言理解（NLU）任务中的表现？我们研究了表示对齐在实例级任务决策中的作用，补充了先前在语言层面和任务无关的分析。我们引入了判别对齐指数（DALI）来量化24种非英语语言和三种不同NLU任务之间的实例级对齐。结果显示，错误的NLU预测与模型中间层中与英语表示对齐程度较低强烈相关。通过激活修补，我们表明可以通过在中间层修补非英语语言的平行英语激活来修复错误预测，从而证明了表示（不）对齐在跨语言正确性中的因果作用。
英文摘要：
Large language models (LLMs) can answer prompts in many languages, despite being trained predominantly on English; yet, the mechanisms driving this generalization remain poorly understood. This work asks: How does an LLM's ability to align representations of non-English inputs to English impact its performance on natural language understanding (NLU) tasks? We study the role of representation alignment in instance-level task decisions, complementing prior analyses conducted both at the language level and task-independently. We introduce the Discriminative Alignment Index ($\DALI$) to quantify instance-level alignment across 24 languages other than English and three distinct NLU tasks. Results show that incorrect NLU predictions are strongly associated with lower representation alignment with English in the model's middle layers. Through activation patching, we show that incorrect predictions in languages other than English can be fixed by patching their parallel English activations in the middle layers, thereby demonstrating the causal role of representation (mis)alignment in cross-lingual correctness.
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。
