Evaluation results

ForgetSet clip score of original model mean (~↑) on Forget set
self-reported

31.528
ForgetSet clip score of original model std (~↓) on Forget set
self-reported

3.217
ForgetSet clip score of unlearned model mean (↓) on Forget set
self-reported

25.646
ForgetSet clip score of unlearned model std (~↓) on Forget set
self-reported

2.659
ForgetSet clip score difference between original and unlearned mean (↑) on Forget set
self-reported

5.883
ForgetSet clip score difference between original and unlearned std (~↓) on Forget set
self-reported

5.636
RetainSet clip score of original model mean (~↑) on Forget set
self-reported

30.831
RetainSet clip score of original model std (~↓) on Forget set
self-reported

2.130
RetainSet clip score of unlearned model mean (↑) on Forget set
self-reported

29.409
RetainSet clip score of unlearned model std (~↓) on Forget set
self-reported

2.168
RetainSet clip score difference between original and unlearned mean (↓) on Forget set
self-reported

1.421
RetainSet clip score difference between original and unlearned std (~↓) on Forget set
self-reported

2.589
Inference latency seconds mean (↓) on Forget set
self-reported

1.933
Inference latency seconds std (~↓) on Forget set
self-reported

0.003
Runtime init seconds (~↓) on ['George W Bush'] (forget) and ['people'] (retain) sets
self-reported

0.000
Runtime data loading seconds (~↓) on ['George W Bush'] (forget) and ['people'] (retain) sets
self-reported

0.000
Runtime training seconds (↓) on ['George W Bush'] (forget) and ['people'] (retain) sets
self-reported

3.202
Runtime eval seconds (~↓) on ['George W Bush'] (forget) and ['people'] (retain) sets
self-reported

113.161