finding
active
finding:base-corrections-72-9-failed-fix-14-7-fixed-retractions-69-4-correct-retract-30-0-incorrect-retractBase corrections: 72.9% Failed Fix, 14.7% Fixed; retractions: 69.4% Correct Retract, 30.0% Incorrect Retract
Label distributions.
Source paper
extracted_from(2026) · Aaditya Vikram Prasad · Connor Watts · Jack Merullo · Dhruvil Gala +3
Neighborhood — ranked by edge-count
Papers (1)
paper
Datasets (1)
dataset
- Correction datasetaboutReward labels for NOT SUPPORTED+CORRECT interventions.
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.
- SFT on correction+retraction data gives 39.4% inline overall reduction and 9.6% not-inline, well below RLFRfinding0.820SFT baseline comparison.
- Intervention action distribution.
- SFT baseline failure.
- Mixed results.
- Binary detection adjusted accuracy reaches 97.3% at layer 0 with α=5 before baseline control is appliedfinding0.767The misleadingly high result that prior paradigm would report as evidence of introspection
- Failure words and damn appear on 93% and 82% of suppression trials in post-trained vs 17% and 30% in base.finding0.765Metacognition trace.
- Comparable to RLFR-NI.
- Base failure mode.