Sharpness-Aware Minimization in Logit Space Efficiently Enhances Direct Preference Optimization

March 18, 2026 ยท Grace Period ยท ๐Ÿ› ICLR 2026

โณ Grace Period
This paper is less than 90 days old. We give authors time to release their code before passing judgment.
Authors Haocheng Luo, Zehang Deng, Thanh-Toan Do, Mehrtash Harandi, Dinh Phung, Trung Le arXiv ID 2603.18258 Category cs.LG: Machine Learning Cross-listed cs.AI Citations 0 Venue ICLR 2026
Abstract
Direct Preference Optimization (DPO) has emerged as a popular algorithm for aligning pretrained large language models with human preferences, owing to its simplicity and training stability. However, DPO suffers from the recently identified squeezing effect (also known as likelihood displacement), where the probability of preferred responses decreases unintentionally during training. To understand and mitigate this phenomenon, we develop a theoretical framework that models the coordinate-wise dynamics in logit space. Our analysis reveals that negative-gradient updates cause residuals to expand rapidly along high-curvature directions, which underlies the squeezing effect, whereas Sharpness-Aware Minimization (SAM) can suppress this behavior through its curvature-regularization effect. Building on this insight, we investigate logits-SAM, a computationally efficient variant that perturbs only the output layer with negligible overhead. Extensive experiments on Pythia-2.8B, Mistral-7B, and Gemma-2B-IT across multiple datasets and benchmarks demonstrate that logits-SAM consistently improves the effectiveness of DPO and integrates seamlessly with other DPO variants. Code is available at https://github.com/RitianLuo/logits-sam-dpo.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Machine Learning