question
active
question:if-a-model-s-features-are-well-calibrated-can-one-use-these-features-as-sources-of-supervisionIf a model's features are well-calibrated, can one use these features as sources of supervision?
Motivating question.
Source paper
extracted_from(2026) · Aaditya Vikram Prasad · Connor Watts · Jack Merullo · Dhruvil Gala +3
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Calibration enables repurposing.
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.
- Model features can serve as inexpensive, dense, scalable supervision for open-ended tasks.claim0.832Central thesis.
- Clamping feature activations causally alters model behavior in interpretable ways.
- Central claim motivating the product.
- features serve as oversight signals to intentionally design models with desirable capabilitiesquote0.775Core vision statement.
- Acknowledged limitation.
- Core goal statement.
- Observation about asymmetry in base model capabilities.
- The central mathematical theorem proved/expounded in the chapter.