List your company

Agent Verification on MIMIC-IV Cholecystitis

0.9648AUROCGLEAN + Active (εu = 0.0)
0.6088080.7012290.793650.886071Mar 3, 2026

Evaluation results

MethodLinks
GLEAN + Active (εu = 0.0)Agent Backbone=Qwen3-30B-A3B-Instruct2026.030.96480.03790.0750.0753
GLEAN + Active (εu = 0.5)Agent Backbone=Qwen3-30B-A3B-Instruct2026.030.95820.0480.06960.0753
GLEAN (K = 3)Agent Backbone=Qwen3-30B-A3B-Instruct2026.030.95590.05050.07180.079
GLEAN + Active (εu = 0.0)Agent Backbone=Qwen2.5-7B-Instruct2026.030.91590.08860.12060.1116
GLEAN + Active (εu = 0.5)Agent Backbone=Qwen2.5-7B-Instruct2026.030.90630.09620.10670.1124
GLEAN (K = 3)Agent Backbone=Qwen2.5-7B-Instruct2026.030.89290.09870.08930.1174
GLEAN (K = 1)Agent Backbone=Qwen3-30B-A3B-Instruct2026.030.88880.11110.08490.1314
GLEAN (K = 1)Agent Backbone=Qwen2.5-7B-Instruct2026.030.85240.10890.06410.1495
Self-ConsistencyAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.84930.14390.10190.1534
Semantic EntropyAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.84480.1566——
ORMAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.83320.15550.19220.2146
ORMAgent Backbone=Qwen2.5-7B-Instruct2026.030.8190.13640.12830.1846
Self-ConsistencyAgent Backbone=Qwen2.5-7B-Instruct2026.030.79780.16460.08340.1484
Self-VerificationAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.79070.20450.14950.2057
RAG-AugmentedAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.78960.18940.25750.2709
Semantic EntropyAgent Backbone=Qwen2.5-7B-Instruct2026.030.77720.1747——
Self-VerificationAgent Backbone=Qwen2.5-7B-Instruct2026.030.75760.17220.2210.2299
RAG-AugmentedAgent Backbone=Qwen2.5-7B-Instruct2026.030.75710.18990.24910.2598
LLM-as-a-JudgeAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.75480.21970.06570.2007
Med-PRMAgent Backbone=Qwen3-30B-A3B-Instruct2026.030.75190.21460.3670.3626
P(TRUE)Agent Backbone=Qwen3-30B-A3B-Instruct2026.030.73840.22220.27580.2927
Med-PRMAgent Backbone=Qwen2.5-7B-Instruct2026.030.64750.25820.3010.3107
LLM-as-a-JudgeAgent Backbone=Qwen2.5-7B-Instruct2026.030.64260.25570.09170.2259
P(TRUE)Agent Backbone=Qwen2.5-7B-Instruct2026.030.62250.2810.3640.3715