WRSWorld Research StateContribute
WORLD RESEARCH STATE

SNLI and MultiNLI: large human-labeled entailment corpora

Contribution grade

Natural-language inference · Evaluation resource design

Reference score160,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Collected entailment, contradiction and neutral hypotheses around text premises; extended caption-based SNLI with multi-genre written and spoken English and matched versus mismatched genre evaluation.

Scientific scope

Collected entailment, contradiction and neutral hypotheses around text premises; extended caption-based SNLI with multi-genre written and spoken English and matched versus mismatched genre evaluation. SNLI workers saw captions, not images, despite scenario grounding. MultiNLI follows the collection template but adds ten genres with only five represented in training. Three-way annotation and genre transfer do not settle unrestricted logical consequence or all-domain language reasoning.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. SNLI workers saw captions, not images, despite scenario grounding. MultiNLI follows the collection template but adds ten genres with only five represented in training. Three-way annotation and genre transfer do not settle unrestricted logical consequence or all-domain language reasoning.

Contributors

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