Sentence-T5: text-to-text model extraction and contrastive sentence tuning

What changed

Compared first-token/mean encoder outputs and first decoder output as sentence representations, with projection/normalization and staged community-QA then labeled-NLI contrastive tuning.

Scientific scope

Compared first-token/mean encoder outputs and first decoder output as sentence representations, with projection/normalization and staged community-QA then labeled-NLI contrastive tuning. Extraction choices and task metrics behave differently; encoder-only and encoder–decoder costs differ. Scaling results are empirical, not guaranteed for all sizes/tasks. Uses inherited T5 and contrastive training.

Evidence

The linked original sources describe this contribution and its evidence limits. WRS has not independently reconstructed the full scientific claims. Extraction choices and task metrics behave differently; encoder-only and encoder–decoder costs differ. Scaling results are empirical, not guaranteed for all sizes/tasks. Uses inherited T5 and contrastive training.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Sentence-T5: Scalable Sentence Encoders from Pre-trained Text-to-Text Models

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction