← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6517.f · 生成模型:GPT-6

步骤奖励与结果奖励分开

理解等级:中级

过程奖励模型评估中间推理步骤,结果验证器通常检查最终答案;通用裁判又可按量规评估更多维度。它们的粒度与训练用途不同,步骤评分也可用于候选选择等场景,不能仅按名称认定能力或质量。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      intermediate steps of reasoning. And unlike verifiers, which check correctness symbolically and usually only at the outcome level, PRMs provide step-by-step reward signals during training in reinforcement learning. We can categorize PRMs as "step-level judges," which are predominantly developed for training, not pure evaluation. (In practice, PRMs are difficult to train reliably at scale. For example, DeepSeek R1 did not adopt PRMs and instead combined

    参考来源

    - 位置:《Build a Reasoning Model (From Scratch)》第10章《附录F 常见模型评估方法》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Build a Reasoning Model (From Scratch)/_epub-src/OEBPS/Text/appendix-f.html) - 结论依据:摘录中关于步骤奖励与结果奖励分开的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:intermediate steps of reasoning. And unlike verifiers, which check correctness symbolically and usually only at the outcome level, PRMs provide step-by-step reward signals during training in reinforcement learning. We can categorize PRMs as "step-level judges," which are predominantly developed for training, not pure evaluation. (In practice, PRMs are difficult to train reliably at scale. For example, DeepSeek R1 did not adopt PRMs and instead combined