quote
active
quote:broadly-turn-averaged-saes-make-interpretability-techniques-practical-at-long-context-lengthsBroadly, turn-averaged SAEs make interpretability techniques practical at long context lengths.
Abstract thesis.
Source paper
extracted_from(2026) · Kevin Der · Harish Kamath · Thompson, Ben
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.
- Abstract.
- Authors' hedged position on SAEs.
- Appendix I.
- Extension of mechanistic interpretability findings to the metacognitive domain
- Forward-looking claim about the potential of model introspection as an interpretability tool
- Claim that feature grounding enables interpretability metrics.
- Assistant-only SAEs reconstruct user turns 1.5–3× worse; both-roles training cuts user-turn nMSE 53–63%finding0.781Appendix G.
- Geometric explanation for prior negative results.