VocalGrad finds audio language models near chance on acoustic change
Human annotators score high on binary temporal-change classification while models stay near chance.
Stories per day · 30 days
2 RSS
2 stories, newest first · one card per story
Stories tagged #linear-probing
Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions
A linear hidden-state direction encodes question impossibility in 1.7B-70B LLMs, but misalignment with the safety-refusal pathway explains why models answer unanswerable questions.