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

Clinical Report Generation on S0001 v1 (test)

48.23BLEU-1

CELM

Updated 3mo ago

Evaluation Results

Method	Links
CELM 2026.01		48.23	55.65	47.34
CELM - SCC 2026.01		33.83	38.43	28.89
Qwen3-4b-instruct-2507 2026.01		27.95	30.38	24.18
Gemma-3-4b-it 2026.01		23.72	28.57	21.47
Meta-llama-3-8b-instruct 2026.01		22.49	26.73	19.46
Llama-3.1-8b-instruct 2026.01		21.28	25.11	18.39
Qwen3-1.7b 2026.01		20.47	20.51	16.83
Llama-3.1-8b-instruct 2026.01		20.4	21.51	18.44
Gemma-3n-e2b-it 2026.01		19.67	24.1	18.75
Llama-3.2-3b-instruct 2026.01		19.51	22.24	18.43
Gemma-3-1b-it 2026.01		18.97	21.12	15.67
Qwen3-4b-instruct-2507 2026.01		18.09	20.58	20.67
Gemma-3n-e4b-it 2026.01		17.93	24.16	18.16
Llama-3.2-3b-instruct 2026.01		16.92	18.44	17.51
Medgemma-4b-it 2026.01		16.18	21	15.29
Medgemma-4b-it 2026.01		15.02	21.96	15.29
Qwen3-1.7b 2026.01		13.35	13.78	11.02
Llama-3.2-1b-instruct 2026.01		3.37	9.19	5.29
Llama-3.2-1b-instruct 2026.01		3.1	8.2	4.73
Gemma-3n-e4b-it 2026.01		1.01	3.97	2.31
Gemma-3-4b-it 2026.01		0.85	0.93	0.78
Gemma-3n-e2b-it 2026.01		0.83	1.16	0.88
Gemma-3-1b-it 2026.01		0.68	0.86	0.64