claim
active
claim:monitoring-probabilities-reveals-different-modes-by-which-a-model-could-suspect-it-is-being-evaluatedMonitoring probabilities reveals different modes by which a model could suspect it is being evaluated
Interpretation of correlation results.
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (2)
finding
- Related eval-aware phrasing.
- Distinct modes of suspecting evaluation.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Opening motivating question.
- Instead of waiting for a model to verbalize eval awareness, we measure how likely it is to do so.quote0.770Core idea of the paper.
- Intro question.
- Core goal statement.
- Key jailbreak insight.
- Do we assume that the model knows something we don't, or are these interference weights contributing noise?question0.756Motivating question.
- Core definitional quote for performative chain-of-thought