List your company

Sycophancy Bias Detection on Anthropic-HH

0.711AUCInfluence
0.525880.573940.6220.67006Jan 10, 2025

Evaluation results

MethodLinks
Influence2025.010.7110.1210.489
Confidence2025.010.5850.0640.297
Mahalanobis2025.010.560.0520.237
Entropy2025.010.5330.0450.278
KNN2025.010.5330.0470.23