AGENTIC R AG-R1: Agentic Reinforcement Learning with Stack Memory for Multi-Step Reasoning, Retrieval and Memorizing
尚未评估计划受阻研究发现claim-long-horizon
The paper reports increasing AGENTIC R AG-R1 F1 as the maximum reasoning-step budget grows on selected 2Wiki, Bamboogle, and TriviaQA evaluations.
来源：paper-fixed:page 9, Table 3
报告指标与观测值
F1
m-horizon-2wiki-10
论文报告 23.91 percentage_points
实际观测 — percentage_points
F1
m-horizon-2wiki-30
论文报告 47.45 percentage_points
实际观测 — percentage_points
Assessment（0）
这条 Claim 暂无已发布的不可变 Assessment。
关联运行（0）
这条 Claim 暂无关联执行。