3 papers
cs.LG2026
Safety-Utility Conflicts Are Not Global: Surgical Alignment via Head-Level Diagnosis
Wang Cai, Yilin Wen, Jinchang Hou +5
Safety alignment in Large Language Models (LLMs) inherently presents a multi-objective optimization conflict, often accompanied by an unintended degradation of general capabilities…
cs.CL2026
SyncThink: A Training-Free Strategy to Align Inference Termination with Reasoning Saturation
Gengyang Li, Wang Cai, Yifeng Gao +1
Chain-of-Thought (CoT) prompting improves reasoning but often produces long and redundant traces that substantially increase inference cost. We present SyncThink, a training-free a…
cs.CL2025
Beyond Demonstrations: Dynamic Vector Construction from Latent Representations
Wang Cai, Hsiu-Yuan Huang, Zhixiang Wang +1
In-Context derived Vector (ICV) methods extract task-relevant representations from large language models (LLMs) and reinject them during inference, achieving comparable performance…