List your company

Agentic Reasoning and Task Execution on Controlled-domain Research

0.314Mean Task Rubric Score (G)Base
0.310880.311690.31250.31331Jul 5, 2026

Evaluation results

MethodLinks
BasePolicy Variant=Base, Seeds=3, Rollouts per task=3, Number of tasks=20, Total trials (N)=1802026.070.314—————
Offline AW ControllerPolicy Variant=AW, Seeds=3, Rollouts per task=3, Number of tasks=20, Total trials (N)=1802026.070.3110.0030.0260.0090.23813.9