Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Positive Movie Review Generation on IMDB (test)
Loading...
0.99
Reward
RRHF-OP-128
0.52096
0.64273
0.7645
0.88627
Apr 11, 2023
Reward
Perplexity
Updated 4d ago
Evaluation Results
Method
Method
Links
Reward
Perplexity
RRHF-OP-128
KL penalty=none
2023.04
0.99
32.081
RRHF
Setting=BP
2023.04
0.861
32.083
RRHF
Setting=B
2023.04
0.799
32.077
PPO
KL penalty=none
2023.04
0.796
42.916
NLPO
KL penalty=none
2023.04
0.777
41.035
RRHF-OP-128
KL penalty=0.1
2023.04
0.635
32.088
PPO
KL penalty=0.1
2023.04
0.626
35.049
NLPO
KL penalty=0.1
2023.04
0.62
34.816
SFT
Setting=baseline
2023.04
0.539
35.472
Feedback
Search any
task
Search any
task