List your company
ResearchFormerly WizWand

Membership Auditing on Math Reasoning Tasks

83.5AUCDIBA
49.07658.01366.9575.887Nov 18, 2025
Updated 5mo ago

Evaluation results

MethodLinks
DIBATarget Algorithm=DAPO, Model Scale=3B2025.1183.515.4
DIBATarget Algorithm=DAPO, Model Scale=7B2025.1181.910.7
DIBATarget Algorithm=GRPO, Model Scale=7B2025.1174.98.4
DIBATarget Algorithm=GRPO, Model Scale=3B2025.1171.37.3
EntropyTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1164.50.8
EntropyTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1160.61.1
EntropyTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1159.80.8
ZlibTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1158.80.2
ZlibTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1158.80.3
EntropyTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Static-target2025.1158.80.5
EntropyTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Static-target2025.1158.60.5
EntropyTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Static-target2025.1158.60.5
EntropyTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Static-target2025.1158.40.2
ZlibTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1158.31.6
LOSSTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Static-target2025.1158.20
EntropyTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1157.90.6
LOSSTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Static-target2025.1157.50.2
LOSSTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1157.10.5
LOSSTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1157.10.2
ZlibTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1157.10.3
LOSSTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Static-target2025.1156.80.6
Min-K%Target Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1156.70.6
LOSSTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Static-target2025.1156.20.8
LOSSTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1155.40.3
LOSSTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1155.30.4
EntropyTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1155.30.9
Min-K%Target Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Static-target2025.1154.80.3
Min-K%Target Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1154.71.6
LOSSTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1154.61.1
ZlibTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Static-target2025.1154.30.1
ZlibTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Static-target2025.1154.10
ZlibTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1154.10.3
ZlibTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Static-target2025.1153.60.2
EntropyTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1153.30.6
EntropyTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1153.30.7
ZlibTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Static-target2025.11530.3
LOSSTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1152.70.7
ZlibTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1152.70.1
Min-K%Target Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1152.60.4
ZlibTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1152.51
Min-K%Target Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1152.50
Min-K%Target Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1152.50.2
Min-K%Target Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Static-target2025.1152.40
Min-K%Target Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1152.40.8
NeighborTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Static-target2025.1152.30.5
Min-K%Target Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1152.20.5
Min-K%Target Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Static-target2025.1152.10.2
ZlibTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1151.90.4
Min-K%Target Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Static-target2025.1151.90.2
NeighborTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Static-target2025.1151.80.4
NeighborTarget Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Static-target2025.1151.80.6
Min-K%Target Algorithm=DAPO, Model Scale=7B, Evaluation Setting=Calibrated2025.1151.70.3
NeighborTarget Algorithm=GRPO, Model Scale=7B, Evaluation Setting=Static-target2025.1151.40.2
LOSSTarget Algorithm=GRPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1151.30.9
EntropyTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Calibrated2025.1151.20.4
LOSSTarget Algorithm=DAPO, Model Scale=3B, Evaluation Setting=Self-generated, Rollouts (N)=82025.1150.40.7