Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Learning to Defer on Synthetic benchmark (test)
Loading...
28.1
Test True Risk
Augmented comp-sum surrogate
26.58
36.84
47.1
57.36
Mar 15, 2026
Test True Risk
Test Excess Risk
Advice Rate
Bayes-Action Match
Updated 1mo ago
Evaluation Results
Method
Method
Links
Test True Risk
Test Excess Risk
Advice Rate
Bayes-Action Match
Augmented comp-sum surrogate
Training size (n)=5000...
2026.03
28.1
0.1
49.6
99.3
Separated
Training size (n)=5000...
2026.03
33.4
5.4
62.5
58.4
L2D
Training size (n)=5000...
2026.03
34.2
6.2
0
48.8
Random route, no advice
Training size (n)=5000...
2026.03
43.2
15.2
0
24.9
Random (j, k)
Training size (n)=5000...
2026.03
55
26.9
50.1
24.9
Learned route, random advice
Training size (n)=5000...
2026.03
66.1
38.1
49.6
24.5
Feedback
Search any
task
Search any
task