concept
active
concept:non-mechanistic-interpretabilityNon-mechanistic interpretability
Explaining activations without understanding mechanism.
Neighborhood — ranked by edge-count
Papers (1)
paper
Frameworks (2)
framework
- Activation OraclesimplementsFramework training LLMs to answer questions about externally-provided activation vectors
- LatentQAassociated_withPrior non-mechanistic approach of training LLMs to answer questions about their own activations; AOs generalize it.
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.
- Functionally significant directions in the residual stream that organize activations in mechanistic interpretability
- The capability to explain model predictions; a central theme of the paper, with disruption profiles as vehicle.
- Proposed paradigm for evaluating interpretability work through empirical falsifiability rather than benchmarks or user studies
- Goodfire's stance drawn from experience with Evo and R1.
- Method using large language models (Claude) to generate and test explanations of features at scale
- Socio-philosophical open question.
- An inference path across features that constitutes a computational subgraph in a trained model
- Advantage of DiffLogic CA over NCA — learned rules are pure binary logic circuits that can be visualized and analyzed