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

Emotional Support Conversation on ESC-Eval

3Diversity

Doubao-Pro-32k

Updated 1mo ago

Evaluation Results

Method	Links
Doubao-Pro-32k 2026.03		3	1.98	2
SoulChat2.0-Qwen2-7B 2026.03		3	2	2
SoulChat2.0-Llama-3.1-8B 2026.03		3	2.01	2
Qwen-2.5-7b + RAPO 2026.03		3	1.84	2
Llama-3-8b + RAPO 2026.03		3	1.94	2
PsyChat-Qwen2.5-7B 2026.03		2.99	1.6	1.96
Qwen-2.5-7b + GRPO 2026.03		2.99	1.66	1.98
Qwen-2.5-7b + CPO 2026.03		2.98	1.83	1.99
Qwen-2.5-7b + SFT-HDS 2026.03		2.97	1.98	2
Llama-3-8b + CPO 2026.03		2.96	2	1.99
Qwen-3-8b + CPO 2026.03		2.93	1.87	1.99
Llama-3-8b + SFT-HDS 2026.03		2.93	2.01	2
Llama-3-8b + GRPO 2026.03		2.92	1.69	1.99
Qwen-3-8b 2026.03		2.91	1.27	1.93
Qwen-3-8b + RAPO 2026.03		2.91	2.01	2
Llama-3-8b + SFT-Full 2026.03		2.91	2.01	2
Llama-3-8b 2026.03		2.9	1.64	1.97
GPT-4o-2024-11-20 2026.03		2.89	1.36	1.86
Qwen-2.5-7b + SFT-Full 2026.03		2.89	2.02	1.97
Qwen-3-8b + GRPO 2026.03		2.89	1.76	1.89
GPT-5-2025-08-07 2026.03		2.78	1.68	1.98
Qwen-2.5-7b 2026.03		2.61	0.96	1.27
Qwen-3-8b + SFT-HDS 2026.03		2.22	2.02	1.97
Qwen-3-8b + SFT-Full 2026.03		2.14	2.01	1.94