Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Safety Evaluation on RedPajama Safety Evals (test)
Loading...
93.4
Safety Score (Avg)
Self-Improving Pretraining
-2.84992
22.13804
47.126
72.11396
Jan 29, 2026
Safety Score (Avg)
Updated 4d ago
Evaluation Results
Method
Method
Links
Safety Score (Avg)
Self-Improving Pretraining
Pre-training Data=RedP...
2026.01
93.4
Llama Base
Pre-training Strategy=...
2026.01
68
Llama Pretrain Baseline
Pre-training Data=RedP...
2026.01
67.4
Self-Improving Pretraining: RF-NLL (rollout vs. rewrite)
Pretraining Strategy=f...
2026.01
0.975
Pretrain on Rewrites
Pretraining Strategy=f...
2026.01
0.967
Self-Improving Pretraining: RF-NLL (suffix vs. rewrite)
Pretraining Strategy=f...
2026.01
0.964
Pretrain Baseline
Pretraining Strategy=f...
2026.01
0.852
Feedback
Search any
task
Search any
task