解析LoRA微调如何让大模型学会判断文本相关性。
How Relevance Emerges: Interpreting LoRA Fine-Tuning in Reranking LLMs
- 通过不同配置的LoRA微调，观察模型如何逐步学习相关性判断。
- 低秩矩阵更新中，注意力层比前馈层对排序精度影响更大。
- 揭示了关键层和投影机制，适合想理解模型内部运作的研究者。
我们对Mistral-7B、LLaMA3.1-8B和Pythia-6.9B在MS MARCO数据集上使用多种LoRA配置进行段落重排序任务的微调，探索了相关性信号如何在训练过程中演化。研究考察了不同LoRA秩（1, 2, 8, 32）的影响，以及更新的多头注意力（MHA）与多层感知机（MLP）组件的相对重要性。消融实验揭示了在LoRA变换中哪些层和投影对重排序准确率最为关键。这些发现为理解LoRA的适应机制提供了新视角，推动信息检索领域更深入的机制研究。本研究使用的全部模型均已公开。
原文摘要 · Abstract (English)
We conduct a behavioral exploration of LoRA fine-tuned LLMs for Passage Reranking to understand how relevance signals are learned and deployed by Large Language Models. By fine-tuning Mistral-7B, LLaMA3.1-8B, and Pythia-6.9B on MS MARCO under diverse LoRA configurations, we investigate how relevance modeling evolves across checkpoints, the impact of LoRA rank (1, 2, 8, 32), and the relative importance of updated MHA vs. MLP components. Our ablations reveal which layers and projections within LoRA transformations are most critical for reranking accuracy. These findings offer fresh explanations into LoRA's adaptation mechanisms, setting the stage for deeper mechanistic studies in Information Retrieval. All models used in this study have been shared.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。