method
active
method:concept-bottleneck-modelsConcept bottleneck models
Intrinsically interpretable models using concepts in training.
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Intrinsic interpretability section.
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.
- Related work designing LLMs to natively support interpretable concept steering
- Norman's design model vs user model; a mental conception of how a system works.
- Central entity of Jackson's framework: a structure invented to give coherent account of immediate consequences of actions; the building block of software design
- Generative models that reverse a noising process, mentioned in quasi-simulator table.
- How a neural network encodes a semantic concept internally, argued to be better captured by manifolds than by atomic features.
- Theme issue context: relates to internal models of environment, central to consciousness and cognition across substrates.
- Learned in pretraining; belief over them is what interventions change.