Hard maximum similarity pooling in late-interaction models induces higher patch-level gradient concentration and greater length sensitivity than top-k or softmax alternatives.
https://www.R-project.org/
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Spike Hijacking in Late-Interaction Retrieval
Hard maximum similarity pooling in late-interaction models induces higher patch-level gradient concentration and greater length sensitivity than top-k or softmax alternatives.