5 papers
CircuitBreak: Pruning a Sparse Safety Circuit of Neurons
Jincheng Ying, Li Wenlin, Wencheng Yang +1
White-box neuron pruning has shown that refusal in large language models (LLMs) can depend on a small, removable set of feed-forward (FFN) channels. Existing attacks either need ca…
Schrödinger--Föllmer Actor--Critic: Diffusion Policy Improvement with Finite-Sample Analysis
Yuling Jiao, Lican Kang, Jerry Zhijian Yang +1
Diffusion policies represent multimodal action distributions, but an advantage-weighted update does not specify how to sample from the resulting target distribution. We propose Sch…
Mask2Shield: Strengthening LLM Safety against Neuron-Pruning Attacks
Ying JinCheng, Minghui Xu, Yinhao Xiao +2
Large language models (LLMs) are safety-aligned before deployment to reduce harmful content generation. Yet neuron-level pruning attacks show that refusal can depend on a small set…
Bridging Diffusion Pruning and Step Distillation with Teacher-Aligned Repair
Jincheng Ying, Li Wenlin, Minghui Xu +1
Diffusion models generate high-quality images, but their inference cost comes from two sources: large denoising networks and repeated denoising steps. Existing compression pipeline…
Efficient Diffusion Distillation via Embedding Loss
Jincheng Ying, Yitao Chen, Li Wenlin +2
Recent advances in distilling expensive diffusion models into efficient few-step generators show significant promise. However, these methods typically demand substantial computatio…