Dialogue Response Generation on DuRecDial benchmark leaderboard · SOTA2 Research

Dialogue Response Generation on DuRecDial

47.43F1LLaMA-3B w/ FF

Evaluation Results

Method	Links
LLaMA-3B w/ FF2026.04		47.43	0.429	0.353	0.006	0.057	57.34	86.4
LLaMA-1B w/ FF2026.04		46.55	0.422	0.349	0.006	0.05	56.35	85.2
LLaMA-3B2026.04		43.19	0.404	0.328	0.006	0.055	50.44	65.1
LLaMA-1B2026.04		42.55	0.397	0.323	0.005	0.047	48.6	61.7
GPT-2 w/ FF2026.04		42.48	0.385	0.308	0.007	0.071	50.47	72.1
TPC-GPT2026.04		41.53	0.379	0.301	0.007	0.075	48.81	74.7
GPT-22026.04		39.78	0.373	0.296	0.007	0.06	45.28	62
DialoGPT w/ FF2026.04		38.65	0.357	0.278	0.006	0.055	45.09	61.5
BART w/ FF2026.04		37.52	0.313	0.241	0.008	0.077	44.49	58.5
TPC-BART2026.04		37.22	0.338	0.255	0.008	0.083	44.52	71.5
DialoGPT2026.04		35.85	0.344	0.266	0.006	0.044	37.95	49.2
BART2026.04		35.22	0.324	0.252	0.007	0.066	39.77	49.9
KERS2026.04		34.04	0.302	0.22	0.005	0.03	40.75	50.5
MGCG_G2026.04		33.48	0.279	0.203	0.007	0.043	35.12	48.6