一作ACL 2026Findings· Findings of ACL
Preventing Safety Drift in Large Language Models via Coupled Weight and Activation Constraints
Songping Peng · Zhiheng Zhang · Daojian Zeng · Lincheng Jiang · Xieping Gao
摘要
Safety alignment in Large Language Models (LLMs) remains highly fragile during fine-tuning, where even benign adaptation can degrade pre-trained refusal behaviors and enable harmful responses. Existing defenses typically constrain either weights or activations in isolation, without considering th...