ScienceQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ScienceQA
96.33Accuracy
39
ScienceQA
88.4Overall Average Score
36
ScienceQA
92.72Accuracy
35
ScienceQA in-scope (test)
0Biology Refusal Count
32
ScienceQA (test)
95.93Accuracy
31
ScienceQA v1.0 (test)
97.47Accuracy (Natural Language Component)
31
ScienceQA
97.07Accuracy
28
ScienceQA
88.6Image Accuracy
26
ScienceQA v1.0 (test)
82.5Accuracy (G1-4)
26
ScienceQA SQA-IMG (test)
91.5SQA-IMG Accuracy
25
ScienceQA
88.45Exact Match (EM)
23
ScienceQA
70.2Accuracy
21
ScienceQA SQA-IMG
74.4SQAIMG Accuracy
21
ScienceQA
92.36AUC
21
ScienceQA SQA-I
69.8Accuracy
21
ScienceQA v1.3 (test)
0.902NAT Score
21
ScienceQA IMG 38
81.3Accuracy
21
ScienceQA
88.7Accuracy
20
ScienceQA
86.81SQA Score
19
ScienceQA (val)
95.3Score
19
ScienceQA
84.2Accuracy
18
ScienceQA Image
100.7Relative Performance
18
ScienceQA Image
94.8Score
17
ScienceQA (SQA^I) (test)
97Accuracy
17
ScienceQA SQA^I v1.0 (test)
74.7SQA^I Score
15