finding
active
finding:no-instruction-baseline-rate-of-target-in-j-lens-is-approximately-zero-across-task-familiesNo-instruction baseline rate of target in J-lens is approximately zero across task families.
Modulation control.
Source paper
extracted_fromAuthors
Neighborhood — ranked by edge-count
Papers (1)
paper
Hypotheses (1)
hypothesis
- Directed modulation prediction.
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.
- Imagine instruction raises property name in J-lens while J-orthogonalized property probe stays at baseline.finding0.819J-space privileged.
- Model organism results.
- Mediation control.
- Related work.
- J-lens eval-awareness score separates NLA-flagged from unflagged positions with AUC 0.853.finding0.785Eval awareness probe.
- Methods comparison.
- Tuned lens skips to output answers and is best next-token predictor while J-lens has highest KL.finding0.779Comparison.
- Limitation.