quote
active
quote:we-were-switching-from-using-saes-to-model-representation-to-modeling-computationWe were switching from using SAEs to model representation to modeling computation.
Pivotal shift that created the faithfulness problem.
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Core motivation.
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.
- Limitation statement.
- Extension of mechanistic interpretability findings to the metacognitive domain
- Authors' hedged position on SAEs.
- Claim that feature grounding enables interpretability metrics.
- Survey of representation engineering methods cited as related work
- Link to prior work.
- Links toy model to practice.
- Broader interpretive claim about LM learning bias inferred from the findings