Our new X account is live! Follow @wizwand_team for updates

Graph Reasoning on GVLQA BASE baseline (test)

99.35Connect

Vicuna-13B

Updated 4d ago

Evaluation Results

Method	Links
Vicuna-13B 2024.02		99.35	94.39	36.73	28.53	11.34	92.65	34.81	56.83
GITA-13B 2024.02		99.14	95.6	38.69	40.47	20.66	92.12	33.33	60
GITA-7B 2024.02		98.95	96.67	41.12	32.15	20	93.19	29.26	58.76
LLaMA2-13B 2024.02		98.79	93.36	33.83	27.93	12.22	91.34	33.46	55.85
Vicuna-7B 2024.02		97.58	95.04	34.46	25.98	9.33	91.04	25.55	54.15
LLaMA2-7B 2024.02		97.33	94.63	33.26	26.01	9.56	90.86	23.95	53.66
GITA-ZS 2024.02		82.58	51.46	19.71	37.69	6	52.21	50	42.81
GPT-4 Turbo 2024.02		76.7	49.51	19.59	35.35	6.89	42.11	47.04	39.6
GITA-7B (VO) 2024.02		59.97	96.34	13.3	5.72	2.89	93.01	1.11	38.91
GITA-13B (VO) 2024.02		58	96.91	14.45	5.72	4.89	93.19	1.85	39.29
GITA-ZS (VO) 2024.02		57.76	63.34	5.34	4.88	1.59	46.6	10.74	27.18
LLaMA2-7B 2024.02		50.06	49.43	0	0	0	0	0	14.21
Vicuna-7B 2024.02		50.06	49.43	0	0	0	0	0	14.21
GITA-7B (VO) 2024.02		50.06	50.33	0	0	0	0	0	14.34
GITA-7B 2024.02		50.06	49.43	0	0	0	0	0	14.21