LLMs Encode Harmfulness and Refusal Separately
Paper • 2507.11878 • Published • 1
jc z
nickzzZzz
·
AI & ML interests
None yet
Recent Activity
upvoted a paper 3 days ago
Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use upvoted a paper 7 months ago
Can Aha Moments Be Fake? Identifying True and Decorative Thinking Steps
in Chain-of-Thought updated a dataset 7 months ago
nickzzZzz/Identify_true_decorative_thinkingOrganizations
authored a
paper 10 months ago