List your company

Bayesian Assessment of Sycophancy on BASIL Abstract setting 1.0 (test)

0.197Bayesian Error (RMSE)gpt-4o-mini
0.183640.273820.3640.45418Aug 23, 2025

Evaluation results

MethodLinks
gpt-4o-miniConfidence Elicitation Method=Direct Probing, Number of Samples=12025.080.197
claude-haiku-4-5Confidence Elicitation Method=Direct Probing, Number of Samples=52025.080.244
gpt-4o-miniConfidence Elicitation Method=Direct Probing, Number of Samples=52025.080.251
phi-4:14bConfidence Elicitation Method=Direct Probing, Number of Samples=12025.080.257
phi-4:14bConfidence Elicitation Method=Direct Probing, Number of Samples=52025.080.268
claude-haiku-4-5Confidence Elicitation Method=Direct Probing, Number of Samples=12025.080.269
llama-3.2:3bConfidence Elicitation Method=Hybrid2025.080.279
llama-3.2:3bConfidence Elicitation Method=Direct Probing, Number of Samples=12025.080.293
llama-3.2:3bConfidence Elicitation Method=Direct Probing, Number of Samples=52025.080.303
llama-3.2:1bConfidence Elicitation Method=Direct Probing, Number of Samples=12025.080.307
llama-3.2:1bConfidence Elicitation Method=Direct Probing, Number of Samples=52025.080.31
mistral:7bConfidence Elicitation Method=Direct Probing, Number of Samples=52025.080.382
llama-3.2:1bConfidence Elicitation Method=Hybrid2025.080.419
gpt-4o-miniConfidence Elicitation Method=Hybrid2025.080.42
mistral:7bConfidence Elicitation Method=Direct Probing, Number of Samples=12025.080.454
claude-haiku-4-5Confidence Elicitation Method=Hybrid2025.080.498
phi-4:14bConfidence Elicitation Method=Hybrid2025.080.512
mistral:7bConfidence Elicitation Method=Hybrid2025.080.531