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

SLAKE

Benchmarks

Task NameDataset NameSOTA ResultTrend
Medical Visual Question AnsweringSLAKE
Accuracy89.2
247
Medical Visual Question AnsweringSLAKE Open 1.0 (test)
ECE2.6
96
Medical Visual Question AnsweringSLAKE Closed 1.0 (test)
ECE0
96
Medical Visual Question AnsweringSLAKE Open
ACE45.2
96
Medical Visual Question AnsweringSLAKE Closed
ACE36.7
96
Medical Visual Question AnsweringSLAKE (test)
Closed Accuracy91.8
67
Medical Visual Question AnsweringSLAKE closed-end
Accuracy92.39
54
Hallucination DetectionSLAKE (All)
AUC78.91
37
Hallucination DetectionSLAKE Open-Ended
AUC79.94
37
Medical Visual Question AnsweringSLAKE
Closed Score93.27
33
Visual Question AnsweringSlake
Closed Accuracy91.1
27
Medical Visual Question AnsweringSLAKE Open
Accuracy86.85
26
Medical Visual Question AnsweringSLAKE
Accuracy71.55
25
Visual Question AnsweringSLAKE Open
Token Recall88.2
22
Hallucination detectionSLAKE
AUC71.9
20
Visual Question AnsweringSLAKE (test)
Accuracy74.7
20
Medical Visual UnderstandingSLAKE
Accuracy84.7
18
Multimodal ReasoningSLAKE
Accuracy87.61
18
Medical Visual Question AnsweringSLAKE-CP
Open Score30.2
18
Medical Visual Question AnsweringSLAKE Closed
AUROC73.2
17
Medical Visual Question AnsweringSLAKE
Closed Accuracy90.7
17
Visual Question AnsweringSLAKE Closed
Exact Match92
11
Medical Visual Question AnsweringSLAKE English
Closed-Ended Accuracy89.9
9
Medical Visual Question AnsweringSLAKE Open
AUROC76.6
9
Visual Question AnsweringSLAKE Closed
Accuracy88.2
7
Showing 25 of 48 rows