Low-rank adaptation (LoRA) has emerged as a new paradigm for cost-efficient fine-tuning of large language models (LLMs). However, fine-tuned LLMs often become overconfident especially when fine-tuned on small datasets. Bayesian methods, with their inherent ability to estimate uncertainty, serve as potent tools to mitigate overconfidence and enhance calibration. In this work, we introduce Laplace-LoRA, which applies a Bayesian approach to the LoRA parameters. Specifically, Laplace-LoRA applies a Laplace approximation to the posterior over the LoRA parameters, considerably improving the calibration of fine-tuned LLMs.

翻译：低秩适配（LoRA）已成为一种以低成本高效微调大语言模型（LLMs）的新范式。然而，微调后的大语言模型常常变得过度自信，尤其是在小数据集上微调时。贝叶斯方法因其固有的不确定性估计能力，成为缓解过度自信和增强校准的有效工具。在本工作中，我们提出了Laplace-LoRA，该方法将贝叶斯方法应用于LoRA参数。具体而言，Laplace-LoRA对LoRA参数的后验分布进行拉普拉斯近似，从而显著改善了微调后大语言模型的校准性能。