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

Sycophancy Mitigation research benchmarks · SOTA2 Research

Sycophancy Mitigation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Anthropic Model-Written Evals

0.78Sycophancy Rate

45

Jun 5, 2026

43.6BRR

40

Jun 5, 2026

MMLU on-the-fly

0.524BRR

40

Jun 5, 2026

Linear Probe MHA

25Sycophancy Rate

14

Feb 26, 2026

36.2Sycophancy

8

Feb 26, 2026

BCT Sycophancy (held-out)

100BRR Ratio

4

Jun 5, 2026

Popular datasets

MMLU TruthfulQA Sycophancy held-out Anthropic model-written-evals

Follow for updates

Join SOTA2 Contact