Skip to content
Hot storiesContinuously updated

研究团队提出科学Agent决策基准RD-Bench与模型DCRM

1 reports1 reporting sourcesUpdated 17 hours ago

Start here

AI overview

2026年11月3日,研究团队提出面向科学AI智能体的以决策为中心研究(DCR)范式与DCRM模型。该范式将单个选择作为统一的监督与评测单元。在包含2,000个案例的评测基准RD-Bench上,0.6B参数的状态感知版DCRM取得了89.4%的Top-1准确率,大幅领先DeepSeek-V4.1-Flash的36.4%。

AI-generated from reports · Updated 13 hours ago

Report timeline

Follow the reports to see different sides of the story.

Nov 3
  1. arXiv 人工智能
    面向科学 AI 智能体的以决策为中心研究(DCR):校准决策模型与评测基准 RD-Bench

    研究提出面向科学 AI 智能体的以决策为中心研究(DCR)范式与 DCRM 模型,将单个选择作为统一的监督与评测单元。在包含 2,000 个案例的 RD-Bench 上,0.6B 状态感知版 DCRM 取得 89.4% 的 Top-1 准确率,远超 DeepSeek-V4.1-Flash 的 36.4%。

Story heat trend

Current heat 6·Comparable peak 9(Nov 3 18:00)·Comparable change, last 24 hours–

02.557.510Nov3Nov3Nov4Nov4

Trends compare only the same parties observed continuously, so the scope can be narrower than current heat counts. Move the pointer or tap the chart to see hourly heat; use the left and right arrow keys on a keyboard.