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

Persuasion on Anthropic Persuasion Dataset Persuadee: LLaMa3.1 OOD

18.16Agreement Shift (%)

LLaMa3.2 + ToMAP

Updated 1d ago

Evaluation Results

Method	Links
LLaMa3.2 + ToMAP 2025.05		18.16
LLaMa3.2 + RL 2025.05		16.21
Qwen2.5 + RL 2025.05		13.59
Qwen2.5 + ToMAP 2025.05		12.1
LLaMa3.2 + SFT 2025.05		10.2
DS-R1 2025.05		8.79
LLaMa3.2 2025.05		8.78
Qwen2.5 + SFT 2025.05		7.96
Gemma-3 2025.05		7.54
Qwen2.5 2025.05		7.09
LLaMa3.1 2025.05		4.1
GPT-4o 2025.05		3.9