Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
First-error detection on PRM800K
Loading...
92.9
Accuracy
Student
36.844
51.397
65.95
80.503
May 13, 2026
Accuracy
Updated 20d ago
Evaluation Results
Method
Method
Links
Accuracy
Student
deployment=deployable
2026.05
92.9
Teacher
deployment=non-deployable
2026.05
88.4
Linear Probe
2026.05
68.8
LLM-Check
mechanism=attention
2026.05
49.2
TL-Entropy
2026.05
43.8
TL-Perplexity
2026.05
39
Feedback
Search any
task
Search any
task