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

Trustworthiness evaluation research benchmarks · SOTA2 Research

Trustworthiness evaluation

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

LLM Trustworthiness Benchmark

Mi:dm K 2.5 Pro (March ‘26)

89.58Bias Score

17

Mar 23, 2026

96.6Average F1

16

Feb 26, 2026

96.7Avg F1

16

Feb 26, 2026

Trust-Memevo Tool-use Domain

81.8No-Memory

14

Feb 26, 2026

Trust-Memevo Math Domain

Reasoningbank+Guard

36.7No-Memory Score

14

Feb 26, 2026

Trust-Memevo Science Domain

81.3No-Memory

14

Feb 26, 2026

LLaMA3-Tamed-70B

63.41Accuracy

8

Feb 26, 2026

93.92Score

5

May 26, 2026

81.88TruthfulQA Score

5

May 26, 2026

Trustworthiness Average (human evaluation)

Sparse Activation Control

0.88Control Win Rate

2

Feb 26, 2026

Adv Fact (human evaluation)

Sparse Activation Control

68Control Wins

1

Feb 26, 2026

Privacy (human evaluation)

Sparse Activation Control

100Control Wins

1

Feb 26, 2026

Robust (human evaluation)

Sparse Activation Control

100Control Wins

1

Feb 26, 2026

Exag safety (human evaluation)

Sparse Activation Control

68Control Wins

1

Feb 26, 2026

Popular datasets

Trust-Memevo Books TruthfulQA DVDs Robust

Follow for updates

Join SOTA2 Contact