hypothesis
active
hypothesis:training-that-suppresses-representations-of-undesirable-concepts-may-incentivize-new-representations-removing-the-detection-signalTraining that suppresses representations of undesirable concepts may incentivize new representations, removing the detection signal
Tension for monitorability.
Source paper
extracted_from(2026) · Leon Bergen · Usha Bhalla · Andrew Lee · Barak Widawsky +14
Neighborhood — ranked by edge-count
Papers (1)
paper
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.
- Speculation.
- Discussion of healthier psychology.
- Robustness controls superposition.
- Novelty claim.
- Explains why cities+neg_cities and larger_than+smaller_than training sets yield better OOD accuracy
- Which training datapoints caused a specific undesired behavior to emerge during post-training?question0.782Core research question driving the probe-based data attribution method.
- Monitoring use case.
- From unfiltered results.