Dialogue Response Generation on DuRecDial 2.0 benchmark leaderboard · SOTA2 Research

Dialogue Response Generation on DuRecDial 2.0

45.84F1LLaMA-3B w/ FF

Evaluation Results

Method	Links
LLaMA-3B w/ FF2026.04		45.84	0.408	0.299	0.018	0.083	67.27	79.8
LLaMA-1B w/ FF2026.04		45.17	0.404	0.297	0.023	0.114	66.91	73.7
GPT-2 w/ FF2026.04		44.74	0.391	0.292	0.017	0.064	64.43	64.8
DialoGPT w/ FF2026.04		44.32	0.382	0.287	0.017	0.062	62.74	65.5
LLaMA-3B2026.04		42.34	0.394	0.288	0.016	0.065	52.8	61.4
GPT-22026.04		41.62	0.375	0.279	0.016	0.06	54.22	50.9
LLaMA-1B2026.04		40.86	0.379	0.271	0.021	0.106	51.47	57.9
BART w/ FF2026.04		40.53	0.306	0.207	0.02	0.071	55.81	67.9
DialoGPT2026.04		39.34	0.359	0.266	0.015	0.054	49.03	40.7
BART2026.04		37.2	0.285	0.186	0.02	0.071	45.27	52
TPC-BART2026.04		36.28	0.296	0.204	0.03	0.093	40.22	63.6
TPC-GPT2026.04		34.62	0.308	0.217	0.025	0.082	38.8	60.7
MGCG_G2026.04		32.26	0.293	0.182	0.016	0.051	29.35	32.2
KERS2026.04		30.11	0.282	0.178	0.017	0.06	33.08	40.6