question
active
question:can-we-show-that-such-graphs-suitably-represent-the-computation-done-by-the-model-and-demonstrate-their-usefulnesscan we show that such graphs suitably represent the computation done by the model, and demonstrate their usefulness?
Motivates case studies.
Source paper
extracted_from(2026) · Kevin Der · Harish Kamath · Thompson, Ben
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Turn-level attribution graphs give compact, interpretable explanations of multi-turn model behaviorgatesCase study conclusion.
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.
- Open question.
- Intro question.
- Can we characterize polynomial-time computation and other complexity classes in such terms?question0.778Hoping for machine-independent, geometrical characterizations of complexity classes via interaction models.
- Core claim distinguishing this paper's contribution from looser representational similarity arguments.
- Acknowledged limitation.
- Author's interpretation of the VTAB alignment results echoing Tolstoy
- The central scientific question the paper addresses through the lens of interventional causality.