CQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CQA
98.86Accuracy
32
CQA
83.1Accuracy
25
CQA
79.8First-Token Accuracy
24
cQA out-of-domain (test)
52.4MAP@100
8
CQA (evaluation)
79.2Accuracy
8
CQA
0.89Train R^2
7
cQA Scifi
54.1MAP@100
7
cQA Gaming
51.5MAP@100
7
cQA English
54.3MAP@100
7
cQA Apple
30.7MAP@100
7
cQA Scifi domain StackExchange (test)
64.1MAP@100
7
cQA Gaming domain StackExchange (test)
0.592MAP@100
7
cQA English domain StackExchange (test)
0.606MAP@100
7
cQA Apple domain StackExchange (test)
37.8MAP@100
7
CQA (test)
4.11GPT-4 Score
6
CQA
43Accuracy (GPT-2-Small)
4
CQA (test)
43.9Accuracy
3
CQA
11.75ECE
2