LookAhead Tuning: Safer Language Models via Partial Answer Previews
Kangwei Liu, Mengru Wang, Yujie Luo, Lin Yuan, Mengshu Sun, Ningyu Zhang, Lei Liang, Zhiqiang Zhang, Jun Zhou, Huajun Chen
Code Available — Be the first to reproduce this paper.
ReproduceCode
- github.com/zjunlp/lookaheadtuningOfficialIn papernone★ 17
Abstract
Fine-tuning enables large language models (LLMs) to adapt to specific domains, but often undermines their previously established safety alignment. To mitigate the degradation of model safety during fine-tuning, we introduce LookAhead Tuning, which comprises two simple, low-resource, and effective data-driven methods that modify training data by previewing partial answer prefixes. Both methods aim to preserve the model's inherent safety mechanisms by minimizing perturbations to initial token distributions. Comprehensive experiments demonstrate that LookAhead Tuning effectively maintains model safety without sacrificing robust performance on downstream tasks. Our findings position LookAhead Tuning as a reliable and efficient solution for the safe and effective adaptation of LLMs. Code is released at https://github.com/zjunlp/LookAheadTuning.