finding
active
finding:on-a-26-920-token-financial-document-turn-averaged-features-surface-macroeconomics-and-subtopics-that-per-token-missesOn a 26,920-token financial document, turn-averaged features surface macroeconomics and subtopics that per-token misses
LongBench v2 example D.
Source paper
extracted_from(2026) · Kevin Der · Harish Kamath · Thompson, Ben
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Long-context results.
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.
- LongBench v2 example E.
- Prediction from coverage results.
- Example B.
- Pairwise coverage.
- §5.5.
- Example A.
- Turn-averaged features have highest embedding similarity 0.663 at k=3; per-token max lowest at 0.617finding0.795Consistent across k=1,2,3,5.
- Example C.