Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Claim improvement suggestion on Subset of 135,828 claims
Loading...
62
Accuracy
FT-DeBERTa
32.256
39.978
47.7
55.422
May 26, 2023
Accuracy
Macro F1
Clarification F1
Typo F1
Links F1
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
Macro F1
Clarification F1
Typo F1
Links F1
FT-DeBERTa
Context=Main thesis
2023.05
62
57.3
65.2
63.1
43.4
FT-DeBERTa
Context=Parent claim
2023.05
60.3
56
63.6
61.2
43
FT-DeBERTa
Context=None
2023.05
59.9
55.4
63.7
60.2
42.5
FT-ELECTRA
Context=Main thesis
2023.05
57.5
52
63.4
54.4
38.3
FT-ELECTRA
Context=Parent claim
2023.05
56.2
50.3
62
53.6
35.3
FT-ELECTRA
Context=None
2023.05
56
49
62.4
52.4
34.5
Random baseline
2023.05
33.4
31.4
38.5
33.4
45.3
Feedback
Search any
task
Search any
task