`包裹，长文用`
`/`
`划分小节。 3. 列表内容用`
- `，引用用`
`，代码用`
`，重要句子可用``。
4. 保留原文所有信息，不得增删改内容。
5. 不要输出``/``/``等文档级标签。
给定的正文内容：
"Decision models answer a question with a typed output, such as a choice among options with a probability for each, so one model can choose an agent's actions, evaluate actions already taken, or score every action to teach a policy with reinforcement learning. Prior work has studied them mainly as actors, judging their usefulness by how well they act. We ask whether a decision model that acts poorly can still teach well. We compare three decision models and three general LLMs as actors and judges on ALFWorld, and use JEV as a teacher on ALFWorld and WebShop. As actors, all six models are pulled away from a subgoal that must come first by an easier one. As judges that see the whole trajectory, the same models credit the steps toward that subgoal, and asking whether each action helped, rather than whether the task will succeed, separates useful from useless actions. As a teacher, JEV raises a Qwen2.5-1.5B policy over GRPO from 72.8% to 82.9% success on ALFWorld and from 56.8% to 71.4% on WebShop. These results suggest that a poor actor can still teach well, because judging an action after the episode ends reveals the consequences that acting must anticipate. Code is available at https://github.com/ZJU-REAL/decision-agents."
分析正文结构：
这是一段连续的学术新闻/摘要文本，没有明显的列表或引用。可以将其划分为几个小节，用``或`
`来划分，或者直接用几个`
`包裹。
考虑到规则2“长文用
`，代码用``，重要句子可用``。 4. 保留原文所有信息，不得增删改内容。 5. 不要输出``/``/``等文档级标签。 给定的正文内容： "Decision models answer a question with a typed output, such as a choice among options with a probability for each, so one model can choose an agent's actions, evaluate actions already taken, or score every action to teach a policy with reinforcement learning. Prior work has studied them mainly as actors, judging their usefulness by how well they act. We ask whether a decision model that acts poorly can still teach well. We compare three decision models and three general LLMs as actors and judges on ALFWorld, and use JEV as a teacher on ALFWorld and WebShop. As actors, all six models are pulled away from a subgoal that must come first by an easier one. As judges that see the whole trajectory, the same models credit the steps toward that subgoal, and asking whether each action helped, rather than whether the task will succeed, separates useful from useless actions. As a teacher, JEV raises a Qwen2.5-1.5B policy over GRPO from 72.8% to 82.9% success on ALFWorld and from 56.8% to 71.4% on WebShop. These results suggest that a poor actor can still teach well, because judging an action after the episode ends reveals the consequences that acting must anticipate. Code is available at https://github.com/ZJU-REAL/decision-agents." 分析正文结构： 这是一段连续的学术新闻/摘要文本，没有明显的列表或引用。可以将其划分为几个小节，用``或`
`来划分，或者直接用几个`
`包裹。 考虑到规则2“长文用
/
划分小节”，我可以将其分为几个主题： 1. 背景与问题 (Decision models answer... can still teach well.) 2. 实验设置 (We compare... WebShop.) 3. 演员与评判者的表现 (As actors... useless actions.) 4. 教师模型的表现与结论 (As a teacher... anticipate.) 5. 代码链接 (Code is available...) 或者，因为原文就是一段，我可以把它拆分成几个自然段，并加上小标题。但原文并没有明显的段落分隔。为了遵守“不得增删改内容”，我最好保持原文的句子，只是用HTML标签包裹。 如果拆分段落，只是把句子组合到不同的`
`中，这不算增删改内容。 让我们尝试划分：
研究背景与问题
Decision models answer a question with a typed output, such as a choice among options with a probability for each, so one model can choose an agent's actions, evaluate actions already taken, or score every action to teach a policy with reinforcement learning. Prior work has studied them mainly as actors, judging their usefulness by how well they act. We ask whether a decision model that acts poorly can still teach well.
实验设置
We compare three decision models and three general LLMs as actors and judges on ALFWorld, and use JEV as a teacher on ALFWorld and WebShop.
演员与评判者的表现
As actors, all six models are pulled away from a subgoal that must come first by an easier one. As judges that see the whole trajectory, the same models credit the steps toward that subgoal, and asking whether each action helped, rather than whether the task will succeed, separates useful from useless actions.
教师模型的表现与结论
As a teacher, JEV raises a Qwen2.5-1.5B policy over GRPO from 72.8% to 82.9% success on ALFWorld and from 56.8% to 71.4% on WebShop. These results suggest that a poor actor can still teach well, because judging an action after the episode ends reveals the consequences that acting must anticipate.
Code is available at
https://github.com/ZJU-REAL/decision-agents”，这里只是一个URL，用``更合适，或者``。规则说代码用``，我可以把URL放在里面。或者直接用``。为了安全，用`https://github.com/ZJU-REAL/decision-agents
`。不过URL通常不算代码，用``就行。规则说“代码用”，我就用``包裹URL吧，或者就普通的``。还是用``严格遵守规则。
等等，原文是 "Code is available at https://github.com/ZJU-REAL/decision-agents."
可以写成：
Code is available at
https://github.com/ZJU-REAL/decision-agents
.
不过``里面通常有换行，单行URL用``更语义化。但规则明确说“代码用”，我就用``。或者把整个句子放进去？不，保留原文格式。
"Code is available at https://github.com/ZJU-REAL/decision-agents." 这样更自然。如果严格要求代码用``，那：
Code is available at https://github.com/ZJU-REAL/decision-agents.
但这改变了段落结构。还是用 `Code is available at
https://github.com/ZJU-REAL/decision-agents
.` 吧。其实 `` 就够了，但为了符合规则，用 ``。
再审视规则：
1. 只输出HTML片段本身，不要Markdown围栏、不要解释；
2. 每个自然段落用包裹，长文用
/
划分小节；
3. 列表内容用- ，引用用
，代码用，重要句子可用；
4. 保留原文所有信息，不得增删改内容；
5. 不要输出//等文档级标签。
划分小节：
背景与问题
Decision models answer a question with a typed output, such as a choice among options with a probability for each, so one model can choose an agent's actions, evaluate actions already taken, or score every action to teach a policy with reinforcement learning. Prior work has studied them mainly as actors, judging their usefulness by how well they act. We ask whether a decision model that acts poorly can still teach well.
实验设计
We compare three decision models and three general LLMs as actors and judges on ALFWorld, and use JEV as a teacher on ALFWorld and WebShop.
实验结果与分析
As actors, all six models are pulled away from a subgoal that must come first by an easier one. As judges that see the whole trajectory, the same models credit the steps toward that subgoal, and asking whether each action helped, rather than whether the task will succeed, separates useful from useless actions.
As a teacher, JEV raises a Qwen2.5-1.5B policy over GRPO from 72.8% to 82.9% success on ALFWorld and from 56.8% to 71.4% on WebShop. These results suggest that a poor actor can still teach well, because judging an action after the episode ends reveals the consequences that acting must anticipate.