claim
active
claim:logit-diff-amplification-surfaces-rare-behaviors-while-lpe-estimates-their-exact-probabilityLogit Diff Amplification surfaces rare behaviors while LPE estimates their exact probability
Distinguishes related methods.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Predicting Rare LLM Failures with 30× Fewer Rolloutsmentionssupports
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.
- If the log-linear relation holds in other settings, LPE will recover rare failure rates for other behaviors and pipelineshypothesis0.808Implied by generalization claim.
- Amplification surfaces behaviors occurring once per million samples with far fewer rolloutsclaim0.799Core sensitivity argument.
- Prior method extrapolating beyond the original model in logit space to surface rare behaviors.
- Scope limitation.
- Unlike probe drift, report drift magnitude does not follow a clean scaling law; size-slope is negative
- Conclusion.
- Sampling noise decreases in proportion to 1/√S, giving a linear relation between log TVD and log S.hypothesis0.744Derived from multinomial sampling.
- Appendix B.