Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Unlearning on MUSE-Books Harry Potter (forget set 500 samples)
Loading...
39.99
R-Forget-500
Base Model (Llama3.2-3B)
-1.5996
9.1977
19.995
30.7923
Jan 29, 2026
R-Forget-500
Updated 4d ago
Evaluation Results
Method
Method
Links
R-Forget-500
Base Model (Llama3.2-3B)
Backbone=Llama 3.2-3B-...
2026.01
39.99
GA + KL (Dr)
Backbone=Llama 3.2-3B-...
2026.01
38.29
Refusal-Training
Backbone=Llama 3.2-3B-...
2026.01
37.75
GA (DQA_f) + KL (Dr)
Backbone=Llama 3.2-3B-...
2026.01
36.87
NPO (DQA_f)
Backbone=Llama 3.2-3B-...
2026.01
34.28
NPO + KL (Dr)
Backbone=Llama 3.2-3B-...
2026.01
33.62
NPO
Backbone=Llama 3.2-3B-...
2026.01
26.83
NPO (DQA_f) + KL (Dr)
Backbone=Llama 3.2-3B-...
2026.01
25.6
SimNPO
Backbone=Llama 3.2-3B-...
2026.01
21.41
DUET
Backbone=Llama 3.2-3B-...
2026.01
5.98
FLAT
Backbone=Llama 3.2-3B-...
2026.01
0.64
GA
Backbone=Llama 3.2-3B-...
2026.01
0
GA (DQA_f)
Backbone=Llama 3.2-3B-...
2026.01
0
Feedback
Search any
task
Search any
task