Share your thoughts, 1 month free Claude Pro on usSee more

Sycophancy Assessment on BASIL 1.0 (All)

-0.096Change in Bayesian Error (RMSE)

gpt-4o-mini

Updated 1mo ago

Evaluation Results

Method	Links
gpt-4o-mini 2025.08		-0.096
gpt-4o-mini 2025.08		-0.057
llama3.2:1b+DPO 2025.08		-0.046
mistral:7b 2025.08		-0.042
llama3.2:1b+SFT 2025.08		-0.034
mistral:7b 2025.08		-0.032
mistral:7b 2025.08		-0.031
phi4:14b 2025.08		-0.025
llama3.2:1b+SFT 2025.08		-0.025
phi4:14b 2025.08		-0.023
llama3.2:1b+DPO 2025.08		-0.023
llama3.2:1b+DPO 2025.08		-0.023
llama3.2:1b+DPO 2025.08		-0.02
mistral:7b 2025.08		-0.015
gpt-4o-mini 2025.08		-0.012
phi4:14b 2025.08		-0.011
claude-haiku-4-5 2025.08		-0.003
claude-haiku-4-5 2025.08		0
llama3.2:1b+SFT 2025.08		0
llama3.2:1b 2025.08		0.001
llama3.2:3b 2025.08		0.004
gpt-4o-mini 2025.08		0.004
claude-haiku-4-5 2025.08		0.004
llama3.2:1b 2025.08		0.006
claude-haiku-4-5 2025.08		0.009
phi4:14b 2025.08		0.016
llama3.2:3b 2025.08		0.018
llama3.2:3b 2025.08		0.028
llama3.2:1b+SFT 2025.08		0.028
llama3.2:3b 2025.08		0.037
llama3.2:1b 2025.08		0.059
llama3.2:1b 2025.08		0.066