Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Zero-shot Natural Language Understanding research benchmarks · SOTA2 Research

Zero-shot Natural Language Understanding

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Standard Zero-shot NLU Suite (ARC-challenge, ARC-easy, BoolQ, HellaSwag, LAMBADA, PIQA, RACE, SciQ, Record, OBQA)

21.22ARC Challenge

18

Feb 26, 2026

LM-Evaluation-Harness ARC, BoolQ, HellaSwag, LAMBADA, PIQA, RACE, SciQ, Record, OBQA

Llama 7B Baseline

46.8ARC Challenge

13

Feb 26, 2026

NLU Benchmark Suite CMNLI, HeSW, PIQA, WSC, CoQA, BoolQ, Race-M, Race-H, XSum, C3

34.43CMNLI Accuracy

8

Apr 23, 2026

65.5Accuracy

8

Feb 26, 2026

Popular datasets

LM Evaluation Harness Standard NLU Suite T0 NLU Benchmark Suite

Follow for updates

Join SOTA2 Contact