English

Bayesian-LoRA: Probabilistic Low-Rank Adaptation of Large Language Models

Artificial Intelligence 2026-04-16 v2

Abstract

Large Language Models usually put more emphasis on accuracy and therefore, will guess even when not certain about the prediction, which is especially severe when fine-tuned on small datasets due to the inherent tendency toward miscalibration. In this work, we introduce Bayesian-LoRA, which reformulates the deterministic LoRA update as a probabilistic low-rank representation inspired by Sparse Gaussian Processes. We identify a structural isomorphism between LoRA's factorization and Kronecker-factored SGP posteriors, and show that LoRA emerges as a limiting case when posterior uncertainty collapses. We conduct extensive experiments on various LLM architectures across commonsense reasoning benchmarks. With only approximately 0.42M additional parameters and 1.2×{\approx}1.2{\times} training cost relative to standard LoRA, Bayesian-LoRA significantly improves calibration across models up to 30B, achieving up to 84% ECE reduction and 76% NLL reduction while maintaining competitive accuracy for both in-distribution and out-of-distribution (OoD) evaluations.

Keywords

Cite

@article{arxiv.2601.21003,
  title  = {Bayesian-LoRA: Probabilistic Low-Rank Adaptation of Large Language Models},
  author = {Moule Lin and Shuhao Guan and Andrea Patane and David Gregg and Goetz Botterweck},
  journal= {arXiv preprint arXiv:2601.21003},
  year   = {2026}
}
R2 v1 2026-07-01T09:24:35.802Z