quote
active
quote:a-harmful-behavior-that-occurs-once-in-a-million-rollouts-will-rarely-surface-during-pre-deployment-testing-yet-will-almost-inevitably-appear-after-releaseA harmful behavior that occurs once in a million rollouts will rarely surface during pre-deployment testing, yet will almost inevitably appear after release.
Motivating statement.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Predicting Rare LLM Failures with 30× Fewer Rolloutscitesmentions
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.
- Caveat on interpretation.
- Main interpretation.
- Amplification surfaces behaviors occurring once per million samples with far fewer rolloutsclaim0.800Core sensitivity argument.
- Steered model's chain-of-thought reasoning concluding it is deployed despite Wood Labs system prompt.
- Concrete EM example.
- Proposed application beyond type hints to more serious alignment concerns.
- Argues against instrumental convergence in GPT.