Showing cs.LGShow all
2 papers · 1 filter
cs.LG2025
Curvature-Aware Safety Restoration In LLMs Fine-Tuning
Thong Bach, Thanh Nguyen-Tang, Dung Nguyen +2
Fine-tuning Large Language Models (LLMs) for downstream tasks often compromises safety alignment, even when using parameter-efficient methods like LoRA. In this work, we uncover a…
cs.LG2025
Rethinking Deep Alignment Through The Lens Of Incomplete Learning
Thong Bach, Dung Nguyen, Thao Minh Le +1
Large language models exhibit systematic vulnerabilities to adversarial attacks despite extensive safety alignment. We provide a mechanistic analysis revealing that position-depend…