List your company

Active Testing on MMBench

68Labeling BudgetActive Testing via Approximate Neyman Allocation
67.5270.767477.24May 11, 2026

Evaluation results

MethodLinks
Active Testing via Approximate Neyman AllocationSurrogate Model=Qwen-3-VL 2B, Target Model=Qwen-3-VL 4B2026.056815
Uniform SamplingSurrogate Model=Qwen-3-VL 2B, Target Model=Qwen-3-VL 4B2026.0580—