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

BioASQ

Benchmarks

Task NameDataset NameSOTA ResultTrend
Hallucination DetectionBioASQ
AUROC81.13
104
Question AnsweringBioASQ
Accuracy98.32
72
Medical Question AnsweringBioASQ
Accuracy88.67
63
Short-form generationBioASQ
AUROC84
32
Selective PredictionBioASQ
E-AURC0.2744
28
Question AnsweringBioASQ (dev)
F1 Score77.8
28
Biomedical reasoningBioASQ out-of-domain
Accuracy91.87
25
Hallucination DetectionBioASQ (test)
AUROC77.51
20
Reliability EstimationBioASQ
AUROC70.46
20
Domain AdaptationBioASQ (test)
BBH54.89
20
Biomedical Multi-hop Question AnsweringBioASQ-B
EM40.6
18
Extractive Question AnsweringBioASQ (test)
EM47.27
16
Snippet RetrievalBIOASQ 7 (test batches 1-5)
MAP0.2518
16
Document RetrievalBIOASQ 7 (test batches 1-5)
MAP19.24
16
Question AnsweringBioASQ MRQA out-of-domain evaluation 2019 (test)
EM60.3
15
Cross-lingual Retrieval-Augmented GenerationBioASQ-ENKB5
ID Score90.99
14
Question AnsweringBioASQ
EM45.68
14
Question AnsweringBioASQ
T Score66
14
Reading ComprehensionBioASQ MRQA out-of-domain
EM67.62
14
Question AnsweringBioASQ factoid 7b (test)
SAcc47.4
13
Hallucination DetectionBioASQ
Inference Throughput (Samples/sec)5,351
12
Extractive Question AnsweringBioASQ MRQA
F1 Score91
12
Biomedical Question AnsweringBioASQ
Factoid Acc29
11
Question AnsweringBioASQ
F1 Score26.1
10
Question AnsweringBioASQ
SAME_CONCLUSION Score85.71
10
Showing 25 of 79 rows