Explanation Generation on TrendFact (test) benchmark leaderboard · SOTA2 Research

Explanation Generation on TrendFact (test)

86.1ECSFactISR(QwQ-32B)

Evaluation Results

Method	Links
FactISR(QwQ-32B)2024.10		86.1	60.1	21.75	78.35	51.56	31.07	45.92
QwQ-32B2024.10		85.31	56.21	19.4	77.12	48.96	28.64	42.47
FactISR(Qwen3-32B)2024.10		84.26	59.26	22.81	79.12	53.65	33.34	48.33
DeepSeek-R1-05282024.10		83.99	56.81	20.41	77.64	50.16	29.54	43.83
Qwen3-32B(Think)2024.10		83.32	55.75	21.81	78.36	51.88	31.38	46.03
QwQ-32B-Preview2024.10		78.43	48.2	15.73	75.25	46.99	27.81	40.48
GPT-4.12024.10		77.26	54.88	22.14	77.35	53.8	33.41	47.36
DeepSeek-V3-03242024.10		76.74	54.27	23.6	79.18	54.23	34.27	48.87
Qwen3-32B(No think)2024.10		75.74	50.01	24.91	80.49	56.49	36.58	51.06
Qwen2.5-72B-instruct2024.10		72.07	50.61	26.32	80.15	56.37	37.12	50.8
DeepSeek-R12024.10		0.9115	—	0.2705	0.8143	0.5832	0.3821	0.5188
o1-preview2024.10		0.8986	—	0.2693	0.8022	0.5602	0.396	0.5206
GPT-4.12024.10		0.8622	—	0.2958	0.827	0.6191	0.4189	0.5561
QwQ-32B-Preview2024.10		0.8198	—	0.2093	0.7804	0.533	0.3459	0.4669
DeepSeek-V32024.10		0.7812	—	0.3573	0.8432	0.6596	0.4805	0.6087
Qwen-72B-instruct2024.10		0.7787	—	0.3366	0.8364	0.6441	0.4589	0.5906