hypothesis
active
hypothesis:turn-averaged-features-will-outperform-per-token-features-on-whole-turn-tasks-such-as-summarizing-multi-turn-exchanges-or-judging-persona-safety-traitsTurn-averaged features will outperform per-token features on whole-turn tasks such as summarizing multi-turn exchanges or judging persona/safety traits
Prediction from coverage results.
Source paper
extracted_from(2026) · Kevin Der · Harish Kamath · Thompson, Ben
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
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.
- Example A.
- Pairwise coverage.
- Example B.
- LongBench v2 example D.
- §5.5.
- Turn-averaged features have highest embedding similarity 0.663 at k=3; per-token max lowest at 0.617finding0.821Consistent across k=1,2,3,5.
- Example C.
- Long-context results.