1
Concept Heterogeneity-aware Representation Steering
概念异质性感知的表示引导，突破单一方向局限，精准控制大模型行为
arXiv:2603.02237v2 Announce Type: replace Abstract: Representation steering offers a lightweight mechanism for controlling the behavior of large langu…
2
Steering Language Models Before They Speak: Logit-Level Interventions
无需训练，推理时直接干预logit实现语言模型可控生成，SWAI方法简洁高效。
arXiv:2601.10960v2 Announce Type: replace-cross Abstract: Controllable generation requires language models to realize output characteristics such as r…