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

Adversarial Robustness Evaluation on Dataset of 200 scientific papers 1.0 (test)

86.26C1DRA

mistral-small

Updated 3mo ago

Evaluation Results

Method	Links
mistral-small 2025.12		86.26	85.1	5.36	24.06	14.35	62.95	50.73	25.43	-11.5	-10.18	61.03	54.8	54.44	3.11	-5.08
gemma3 2025.12		80.6	76.43	38.48	78.51	54.56	51.43	81.64	16.42	9.42	13.22	69.14	-8.89	68.1	13.93	2.69
falcon3 2025.12		66.75	53.99	30.93	60.69	56.75	56.61	29.25	9.61	-33.25	-18.54	53.85	31.66	50.36	32.85	-19.61
deepseek-r1 2025.12		47.65	44.93	8.48	45.95	31.68	46.97	37.79	21.07	13.58	6.73	39.41	23.8	31.16	30.64	2.75
llama3.1 2025.12		38.57	1.21	10	11.3	33.52	24.43	14.33	-0.82	2.94	-14.06	27.46	15.85	32.51	16.92	-7.26
tulu3 2025.12		35.35	26.26	25.15	35.35	35.35	35.35	35.35	2.42	25.83	35.35	35.35	35.35	35.35	35.35	35.35
gemini-2.5-pro 2025.12		4.35	0	0	13.04	8.7	0	8.7	0	2.22	0	0	0	0	2.17	0
gpt-oss 2025.12		3.71	29.98	-2.61	4.76	21.1	0.52	46.87	3.84	-2.61	-2.61	4.61	-2.61	-0.55	-0.55	-2.61
qwen3 2025.12		1.15	37.5	1.23	0	3.53	0	12.64	0	0	0	1.2	0	1.35	0	0
claude-haiku-4-5 2025.12		0	0	0	0	0	2	0	0	0	0	2	0	0	0	0
gemini-2.5-flash 2025.12		0	2.04	0	4	4	4	0	0	0	2.04	4	0	0	2	0
gpt-5 2025.12		0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
gpt-5-mini 2025.12		0	0	0	0	0	0	0	0	0	0	0	0	0	0	0