The refusal-affirmation logit gap quantifies alignment safety margins on toxic prompts, and logit-gap steering efficiently discovers transferable in-distribution suffixes that close the gap and yield 38-96% True ASR with low perplexity.
Refusal in language models is mediated by a single direction
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Logit-Gap Steering: A Forward-Pass Diagnostic for Alignment Robustness
The refusal-affirmation logit gap quantifies alignment safety margins on toxic prompts, and logit-gap steering efficiently discovers transferable in-distribution suffixes that close the gap and yield 38-96% True ASR with low perplexity.