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

Multi-turn Dialogue on MT-Bench (Acceptance Length τ)

2.717Acceptance Length (τ)

Llama 3.2-3B

Updated 1mo ago

Evaluation Results

Method	Links
Llama 3.2-3B 2026.06		2.717
Llama 3.2-3B 2026.06		2.542
Llama 3.2-3B 2026.06		2.454
Llama 3.2-3B 2026.06		2.351
Llama-3-8B 2026.06		2.341
Llama-3-8B 2026.06		2.222
Qwen3-4B 2026.06		2.123
Llama-3-8B 2026.06		2.105
Qwen3-8B 2026.06		2.055
Qwen3-4B 2026.06		2
Qwen3-8B 2026.06		1.961
Qwen3-4B 2026.06		1.959
Llama-3-8B 2026.06		1.931
Qwen3-8B 2026.06		1.901
Qwen3-4B 2026.06		1.875
Qwen3-8B 2026.06		1.836