Steer2Adapt: Dynamically Composing Steering Vectors Elicits Efficient Adaptation of LLMs
Paper • 2602.07276 • Published • 11
Pengrui Han
barryhpr
AI & ML interests
None yet
Organizations
upvoted 6
papers 2 months ago
In-Context Learning May Not Elicit Trustworthy Reasoning: A-Not-B Errors in Pretrained Language Models
Paper • 2409.15454 • Published • 2
The Personality Illusion: Revealing Dissociation Between Self-Reports & Behavior in LLMs
Paper • 2509.03730 • Published • 2
Large Language Model Reasoning Failures
Paper • 2602.06176 • Published • 12
SocialVeil: Probing Social Intelligence of Language Agents under Communication Barriers
Paper • 2602.05115 • Published • 20
Good SFT Optimizes for SFT, Better SFT Prepares for Reinforcement Learning
Paper • 2602.01058 • Published • 43
upvoted a paper 4 months ago