Natural Language Processing research benchmarks

Benchmarks

Dataset NameSOTA methodMetricTrendResultsLast Updated

Hebrew NLP Benchmarks

Hebatron

91.2SNLI Accuracy

May 13, 2026

Federated Dataset Test-Time Personalization 2

FedDPA-F

71.64Paraphrasing

Feb 26, 2026

Federated Dataset 1 Test-Time Personalization

FedDPA-F

78.1Paraphrase Accuracy

Feb 26, 2026

12 NLP task categories Average across distributions

FlexLoRA

1.56Avg Improvement (%)

Feb 26, 2026

12 NLP task categories Heavy-Tail (S) distribution

FlexLoRA

0.017Avg Percentage Improvement

Feb 26, 2026

12 NLP task categories Normal distribution

FlexLoRA

1.34Avg Percentage Improvement

Feb 26, 2026

12 NLP task categories Heavy-Tail (L) distribution

FlexLoRA

1.24Average Improvement

Feb 26, 2026

12 NLP task categories Uniform distribution

FlexLoRA

1.99Avg Improvement

Feb 26, 2026

STS-B (test)

FFT

89.52Accuracy

Jun 15, 2026

DABS Eng Text 1.0 (downstream)

ShED

48.4Accuracy (%)

Feb 26, 2026

CrossFit (test)

Multi-task Learning (MTL)

1GLUE CoLA Score

Feb 26, 2026

E2E (test)

FFT

2.98Perplexity

Jun 15, 2026