List your company

ResearchFormerly WizWand

Sentiment Analysis

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

BERT fine-tuned

93.3F1 Score

1

Feb 26, 2026

Amazon varying Spurious levels with scaled noise intensity (test)

—

—Primary metric

0

May 14, 2026

Amazon varying Spurious levels: 90%, 70%, 50%, 30%, 10% (test)

—

—Primary metric

0

May 14, 2026

Yelp varying Spurious levels: 90%, 70%, 50%, 30%, 10% (test)

—

—Primary metric

0

May 14, 2026

human-sentiment WUDA 20% noise rate

—

—Primary metric

0

Feb 26, 2026

Human-sentiment WUDA target-domain 45% noise rate

—

—Primary metric

0

Feb 26, 2026