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

Fidelity Evaluation on BDI/E Dialogue Dataset

93B-M

Qwen3-14B

Updated 26d ago

Evaluation Results

Method	Links
Qwen3-14B 2026.06		93	80	51	11.2	95.3
GPT-5.5 2026.06		81	79	27	13.3	96.2
Dual-LLM 2026.06		81	59	28	15.8	96
DSv4-Pro 2026.06		79	66	32	17.2	96.5
CogWM-14B 2026.06		39	26	14	32.2	94.8