finding
active
finding:apd-fails-to-learn-ground-truth-mechanisms-in-tms-with-hidden-identityAPD fails to learn ground-truth mechanisms in TMS with hidden identity
New models harder for APD.
Source paper
extracted_from(2025) · Lucius Bushnaq · Dan Braun · Lee Sharkey
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Central claim.
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.
- Robustness.
- SPD avoids shrinkage.
- Ground truth mechanisms identified.
- Replication of Fontana et al. 2025 findings in the paper's own Experiment 2 baseline condition
- Implication of PRH for language model visual grounding
- Tested in Section 4.4 calibration experiment; confirmed by findings.
- Core claim about why pernicious divergence undermines mechanistic conclusions
- Establishes task difficulty as a hard limit that instructions cannot overcome.