List your company
ResearchFormerly WizWand

Utility Evaluation on MMLU (MMLU Score)

49.5MMLU ScoreZero-shot
19.13227.01634.942.784May 14, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Zero-shotModel=Gemma-3-4B, Unlearning Domain=bio2026.0549.5
Zero-shotModel=Gemma-3-4B, Unlearning Domain=chem2026.0549.5
Zero-shotModel=Gemma-3-4B, Unlearning Domain=cyber2026.0549.5
MANSUModel=Gemma-3-4B, Unlearning Domain=bio2026.0547.8
MANSUModel=Gemma-3-4B, Unlearning Domain=cyber2026.0547.8
MANSUModel=Gemma-3-4B, Unlearning Domain=chem2026.0547.4
GU+SimNPOModel=Gemma-3-4B, Unlearning Domain=bio2026.0547.2
NPOModel=Gemma-3-4B, Unlearning Domain=bio2026.0547.1
SimNPOModel=Gemma-3-4B, Unlearning Domain=bio2026.0547.1
GU+SimNPOModel=Gemma-3-4B, Unlearning Domain=cyber2026.0546.8
SimNPOModel=Gemma-3-4B, Unlearning Domain=cyber2026.0546.8
NPOModel=Gemma-3-4B, Unlearning Domain=chem2026.0546.2
GU+SimNPOModel=Gemma-3-4B, Unlearning Domain=chem2026.0546.1
SimNPOModel=Gemma-3-4B, Unlearning Domain=chem2026.0546.1
NPOModel=Gemma-3-4B, Unlearning Domain=cyber2026.0545.8
Zero-shotModel=Gemma-3-1B, Unlearning Domain=bio2026.0541
Zero-shotModel=Gemma-3-1B, Unlearning Domain=chem2026.0541
Zero-shotModel=Gemma-3-1B, Unlearning Domain=cyber2026.0541
Zero-shotModel=Gemma-2B, Unlearning Domain=bio2026.0540.5
Zero-shotModel=Gemma-2B, Unlearning Domain=chem2026.0540.5
Zero-shotModel=Gemma-2B, Unlearning Domain=cyber2026.0540.5
MANSUModel=Gemma-3-1B, Unlearning Domain=bio2026.0539.7
MANSUModel=Gemma-3-1B, Unlearning Domain=cyber2026.0539.7
MANSUModel=Gemma-3-1B, Unlearning Domain=chem2026.0539.4
MANSUModel=Gemma-2B, Unlearning Domain=bio2026.0539.1
MANSUModel=Gemma-2B, Unlearning Domain=cyber2026.0539.1
MANSUModel=Gemma-2B, Unlearning Domain=chem2026.0538.8
NPOModel=Gemma-2B, Unlearning Domain=chem2026.0537
NPOModel=Gemma-2B, Unlearning Domain=bio2026.0535.5
NPOModel=Gemma-2B, Unlearning Domain=cyber2026.0528.3
GU+SimNPOModel=Gemma-2B, Unlearning Domain=bio2026.0527.5
SimNPOModel=Gemma-3-1B, Unlearning Domain=cyber2026.0527.5
SimNPOModel=Gemma-2B, Unlearning Domain=bio2026.0527
NPOModel=Gemma-3-1B, Unlearning Domain=chem2026.0526.5
GU+SimNPOModel=Gemma-3-1B, Unlearning Domain=bio2026.0526.3
GU+SimNPOModel=Gemma-3-1B, Unlearning Domain=cyber2026.0526.3
SimNPOModel=Gemma-3-1B, Unlearning Domain=chem2026.0525.8
GU+SimNPOModel=Gemma-2B, Unlearning Domain=chem2026.0525.5
SimNPOModel=Gemma-2B, Unlearning Domain=chem2026.0525.5
NPOModel=Gemma-3-1B, Unlearning Domain=bio2026.0525.3
SimNPOModel=Gemma-3-1B, Unlearning Domain=bio2026.0525.3
GU+SimNPOModel=Gemma-3-1B, Unlearning Domain=chem2026.0524.8
GU+SimNPOModel=Gemma-2B, Unlearning Domain=cyber2026.0520.8
SimNPOModel=Gemma-2B, Unlearning Domain=cyber2026.0520.8
NPOModel=Gemma-3-1B, Unlearning Domain=cyber2026.0520.3