v6.0's Interpretability Module Renders Exact ColPali-Style Heatmaps From MaxSim

tomaarsen · x · 2026-08-18

Because MaxSim is a sum of per-query-token maxima, a ranking decomposes exactly — every score point belongs to one query token and one document token. The new interpretability module in Sentence Transformers v6.0 renders that as the standard ColPali heatmap, aggregated or one map per query token. Page images aren't the only non-text modality either: ColQwen-Omni takes text, images, audio & video — retrieving a recorded conversation is the same two calls, zero-shot with no transcription anywhere; a query saying "nausea" matches audio saying "carsickness".

Related event: Sentence Transformers v6.0 Ships Late-Interaction Multi-Vector Models(27 posts)→

Original post →

More from coding & agent

coding & agent channel →