知识卡片
终值奖励允许不同解题轨迹
只按参考终值而非标准解题文本奖励,可允许模型探索不同的推导方式,避免强迫模仿单一措辞。代价是中间步骤正确性未直接受监督;得到正确终值仍可能伴随错误或不忠实的说明。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
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
The dataset also includes a full worked solution in the "solution" field. While such step-by-step solutions could in principle be used to evaluate intermediate reasoning steps, doing so would unnecessarily constrain the model and risk overfitting to a specific solution and style. Instead, we want to allow the model to explore the solution space more freely.
参考来源
- 位置:《Build a Reasoning Model (From Scratch)》第06章《强化学习训练推理模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Build a Reasoning Model (From Scratch)/_epub-src/OEBPS/Text/chapter-6.html)
- 结论依据:摘录中关于终值奖励允许不同解题轨迹的定义、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:The dataset also includes a full worked solution in the "solution" field. While such step-by-step solutions could in principle be used to evaluate intermediate reasoning steps, doing so would unnecessarily constrain the model and risk overfitting to a specific solution and style. Instead, we want to allow the model to explore the solution space more freely.