Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Reward rollout alignment on 10-task benchmark T2: Cleaning Whiteboard
Loading...
0.667
Rollout ρ
SARM2
-0.48844
-0.18847
0.1115
0.41147
Jun 9, 2026
Rollout ρ
Updated 1mo ago
Evaluation Results
Method
Method
Links
Rollout ρ
SARM2
standard_full=true
2026.06
0.667
SARM2
Spectral Embedding=exc...
2026.06
0.556
SARM2
Multi-Goal=excluded
2026.06
0.333
RW
2026.06
0.222
Robometer
Fine-tuning=true
2026.06
0
Robometer
Fine-tuning=false
2026.06
-0.111
TR
2026.06
-0.444
Feedback
Search any
task
Search any
task
