claim
active
claim:the-role-of-interpretability-is-not-only-to-decompose-representations-into-atoms-but-to-infer-what-the-atoms-must-beThe role of interpretability is not only to decompose representations into atoms but to infer what the atoms must be.
Closing position.
Source paper
extracted_from(2026) · Thomas Fel · Matthew Kowal · Mozes Jacobs · Dron Hazra +21
Neighborhood — ranked by edge-count
Papers (1)
paper
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.
- Motivation for VPD's parameter-focused approach.
- Socio-philosophical open question.
- Forward-looking suggestion for how inner interpretability could extend the indicator method
- Motivates shift from studying model activations ('thoughts') to understanding parameters ('the computations themselves').
- Benchmark discipline.
- Interpretability hypotheses are often conflated with conclusions and require extensive validationclaim0.784Validation is under-done.
- Proposed paradigm for evaluating interpretability work through empirical falsifiability rather than benchmarks or user studies
- Motivation for the release.