Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Text Classification on AG News (Accuracy and Robustness Metrics)
Loading...
36.42
Accuracy (Avg)
Task Objective
24.5432
27.6266
30.71
33.7934
May 8, 2026
Accuracy (Avg)
Accuracy (Unseen)
Accuracy (OOD)
Path Sensitivity (Logit)
Path Sensitivity (Hidden Layer)
Expected Calibration Error (ECE)
Updated 22d ago
Evaluation Results
Method
Method
Links
Accuracy (Avg)
Accuracy (Unseen)
Accuracy (OOD)
Path Sensitivity (Logit)
Path Sensitivity (Hidden Layer)
Expected Calibration Error (ECE)
Task Objective
Objective=task
2026.05
36.42
36.39
27.53
0.047
0.1387
0.0073
Reveal-path Objective
Objective=reveal
2026.05
28.04
28.02
26.03
12.0342
4.0788
0.0337
Full Field Objective
Objective=full
2026.05
25
25
25.05
0.1302
0.1644
0.2182
Feedback
Search any
task
Search any
task