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

Measurement Information Extraction on MeasEval (test)

51.9Overall Score

LIORI

Updated 1mo ago

Evaluation Results

Method
LIORI 2026.04	51.9	86.1	72.2	64.2	43.7	48.2	46.7	31.8	16.3	9.2
LIORI 2026.04	51.9	86.1	72.2	64.2	43.7	48.2	46.7	31.8	16.3	9.2
MeasHalu 2026.04	51.2	84.8	86	60.7	45.5	47.2	44.2	31	17	10
MeasHalu-7B 2026.04	51.2	84.8	86	60.7	45.5	47.2	44.2	31	17	10
MeasHalu 2026.04	47.9	84.6	86.3	61	42.9	43.3	39.7	27.2	15.5	6.3
MeasHalu-7B 2026.04	47.9	84.6	86.3	61	42.9	43.3	39.7	27.2	15.5	6.3
CONNER 2026.04	47.3	85.5	71.9	52.3	39.8	42.4	43.7	25.7	0	0
CONNER 2026.04	47.3	85.5	71.9	52.3	39.8	42.4	43.7	25.7	0	0
MeasHalu 2026.04	46	81	87.5	44	39.6	41.8	40.7	27.7	11	4.9
MeasHalu-3B 2026.04	46	81	87.5	44	39.6	41.8	40.7	27.7	11	4.9
Gemini-2.5-Pro 2026.04	44	78.2	88.2	48.6	43.6	37.6	38.6	28	14.3	5.6
Gemini-2.5-Pro 2026.04	44	78.2	88.2	48.6	43.6	37.6	38.6	28	14.3	5.6
MeasHalu 2026.04	43.3	78.2	85	44.9	37	37.7	36.5	24.5	8.4	4.3
MeasHalu-3B 2026.04	43.3	78.2	85	44.9	37	37.7	36.5	24.5	8.4	4.3
Counts 2026.04	43.2	86.1	80.4	61.4	40.6	31.1	24.5	18.3	7.7	6.4
Counts 2026.04	43.2	86.1	80.4	61.4	40.6	31.1	24.5	18.3	7.7	6.4
GPT-5 2026.04	40.6	72.4	81.7	50	39.7	35.1	35.5	22.6	13.8	4.2
GPT-5 2026.04	40.6	72.4	81.7	50	39.7	35.1	35.5	22.6	13.8	4.2
GPT-5 2026.04	37.1	80.4	74.2	39.5	36.1	27	35.5	2	15.2	5.2
GPT-5 2026.04	37.1	80.4	74.2	39.5	36.1	27	35.5	2	15.2	5.2
Gemini-2.5-Pro 2026.04	35.9	71.2	78.4	46.4	30.6	26.6	28.7	9	14.6	7.6
Gemini-2.5-Pro 2026.04	35.9	71.2	78.4	46.4	30.6	26.6	28.7	9	14.6	7.6
MeasHalu 2026.04	34.7	65.9	76	26.8	28.4	27.9	28.7	18.4	5.9	3.7
MeasHalu-0.5B 2026.04	34.7	65.9	76	26.8	28.4	27.9	28.7	18.4	5.9	3.7
DeepSeek-V3 2026.04	32	56.7	72.6	35.5	30.3	22.5	22.6	14.9	1.9	0
DeepSeek-V3 2026.04	32	56.7	72.6	35.5	30.3	22.5	22.6	14.9	1.9	0
MeasHalu 2026.04	31.2	64.9	71.7	23.9	26.3	24.1	23.8	14	6.9	2.2
MeasHalu-0.5B 2026.04	31.2	64.9	71.7	23.9	26.3	24.1	23.8	14	6.9	2.2
DeepSeek-R1 2026.04	30.4	58.9	71.1	35.6	26	19.8	18.2	11.8	11.3	5.7
DeepSeek-R1 2026.04	30.4	58.9	71.1	35.6	26	19.8	18.2	11.8	11.3	5.7
Qwen2.5-72b-inst 2026.04	28.6	64.4	82.6	23.6	19.6	14.7	16.4	0.1	7.6	2.1
Qwen2.5-72b-inst 2026.04	28.6	64.4	82.6	23.6	19.6	14.7	16.4	0.1	7.6	2.1
DeepSeek-V3 2026.04	27.1	65.7	76.8	21.4	23.9	13.5	11.3	0.1	8.5	1.4
DeepSeek-V3 2026.04	27.1	65.7	76.8	21.4	23.9	13.5	11.3	0.1	8.5	1.4
DeepSeek-R1 2026.04	25.3	56.9	58.6	24	21.6	16.3	16.3	2.4	8.5	2.9
DeepSeek-R1 2026.04	25.3	56.9	58.6	24	21.6	16.3	16.3	2.4	8.5	2.9
Baseline 2026.04	23.9	82.7	56.1	0	5.3	7.5	6.4	0.7	0.5	0
Baseline 2026.04	23.9	82.7	56.1	0	5.3	7.5	6.4	0.7	0.5	0
Qwen2.5-72b-inst 2026.04	21.2	40.3	51.6	23.2	20.4	13.7	11.3	8.7	3.8	1.2
Qwen2.5-72b-inst 2026.04	21.2	40.3	51.6	23.2	20.4	13.7	11.3	8.7	3.8	1.2
Qwen2.5-7b-inst 2026.04	17.1	49.1	47.8	10.6	8.8	4.5	5.7	0	4	1.7
Qwen2.5-7b-inst 2026.04	17.1	49.1	47.8	10.6	8.8	4.5	5.7	0	4	1.7
Qwen2.5-7b-inst 2026.04	7.3	15.1	16	2.7	6.6	5.9	4.4	3.1	0.3	0.5
Qwen2.5-7b-inst 2026.04	7.3	15.1	16	2.7	6.6	5.9	4.4	3.1	0.3	0.5