包裹，所以整个正文先作为一个
吗？不对，原文里有没有可以划分小节的地方？比如“To improve these signals...”后面讲方法，“Across 9B, 27B...”后面讲实验结果，可能需要用
划分？不过原文是一段，但规则说长文用
/
划分小节。那看看原文的内容，是否有逻辑上的小节划分。 原文第一句：Social media translation...However, gaps...To improve these signals, we use...A calibrated...The Agent...Across 9B...Human evaluation...Retrieval improves...Training ablations... 可能的逻辑划分：第一部分讲背景和问题（前两句），第二部分讲方法（To improve...The Agent...第三到五句），第三部分讲实验结果（Across...最后三句）。所以可以用
来划分方法、结果部分？ 看规则里的例子，引用用blockquote，列表用ul，代码用pre，重要句子用strong。原文里没有列表和代码，所以主要处理段落和标题。 首先，第一部分背景和问题：“Social media translation requires preserving meaning while capturing the cultural nuances of user-generated content (UGC). However, gaps in the cultural knowledge of LLM-based evaluators can lead to inaccurate assessments and unreliable training rewards.” 这部分作为第一个
？ 然后“To improve these signals, we use retrieval-enhanced cultural evaluation to train CUE-MT, a family of social media translation models. A calibrated, domain-specific Judger guides supervised data refinement and is augmented by an Agent for hard-example mining and reinforcement learning. The Agent retrieves relevant external evidence as needed and distills it into concise hints that the Judger uses to score translations and assign rewards.” 这是方法部分，用
？ 接下来“Across 9B, 27B, and 35B-A3B backbones, CUE-MT outperforms the evaluated open-source baselines of comparable size on CUE-MT-Bench and three public benchmarks. Human evaluation on CUE-MT-Bench shows cultural effectiveness gains of 7.44-15.12 percentage points over the strongest external baseline in each size group. Retrieval improves the Judger's four-way classification accuracy by 8.07 percentage points over a matched no-retrieval baseline that uses LLM-generated hints alone. Training ablations further support the combination of hard-example mining and Agent-enhanced rewards.” 这是结果部分，用
？ 然后检查是否有重要句子需要用，比如CUE-MT的全称？原文里“CUE-MT, a family of social media translation models”，可能不需要，但如果有关键术语可以加，但原文没有特别标注，所以可能不用。 现在开始排版： 首先背景部分：
Social media translation requires preserving meaning while capturing the cultural nuances of user-generated content (UGC). However, gaps in the cultural knowledge of LLM-based evaluators can lead to inaccurate assessments and unreliable training rewards.
然后方法部分，用
，里面的句子是否要分段落？原文方法部分是三句话，属于一个自然段，所以用
包裹整个方法内容？不对，
是标题，所以应该是
后面跟内容。比如：
To improve these signals, we use retrieval-enhanced cultural evaluation to train CUE-MT, a family of social media translation models.
A calibrated, domain-specific Judger guides supervised data refinement and is augmented by an Agent for hard-example mining and reinforcement learning. The Agent retrieves relevant external evidence as needed and distills it into concise hints that the Judger uses to score translations and assign rewards.
不对，可能更好的是，方法部分作为一个
标题，然后内容在
里。比如：
Method
？但原文没有“Method”这个词，规则说不能增删内容，所以不能用原文没有的标题。所以必须用原文的句子作为标题？或者原文的逻辑划分直接用包裹相关句子？ 规则里说“长文用
/
划分小节”，所以小节标题应该来自原文的关键句子。比如原文中“To improve these signals, we use retrieval-enhanced cultural evaluation to train CUE-MT, a family of social media translation models.” 这句话是方法的核心，可以作为
。 然后“A calibrated, domain-specific Judger guides supervised data refinement and is augmented by an Agent for hard-example mining and reinforcement learning. The Agent retrieves relevant external evidence as needed and distills it into concise hints that the Judger uses to score translations and assign rewards.” 这两句作为
。 然后结果部分的第一句“Across 9B, 27B, and 35B-A3B backbones, CUE-MT outperforms the evaluated open-source baselines of comparable size on CUE-MT-Bench and three public benchmarks.” 作为
？ 然后后面的三句作为
？ 或者看原文的结构，是否是： 1. 背景：前两句（
） 2. 方法：接下来的三句（
开头，
内容） 3. 结果：最后四句（
开头，
内容） 现在按照规则，不能增删改内容，所以标题必须是原文的句子。比如：
Social media translation requires preserving meaning while capturing the cultural nuances of user-generated content (UGC). However, gaps in the cultural knowledge of LLM-based evaluators can lead to inaccurate assessments and unreliable training rewards.
To improve these signals, we use retrieval-enhanced cultural evaluation to train CUE-MT, a family of social media translation models.
A calibrated, domain-specific Judger guides supervised data refinement and is augmented by an Agent for hard-example mining and reinforcement learning. The Agent retrieves relevant external evidence as needed and distills it into concise hints that the Judger uses to score translations and assign rewards.
Across 9B, 27B, and 35B-A3B backbones, CUE-MT outperforms the evaluated open-source baselines of comparable size on CUE-MT-Bench and three public benchmarks.
Human evaluation on CUE-MT-Bench shows cultural effectiveness gains of 7.44-15.12 percentage points over the strongest external baseline in each size group. Retrieval improves the Judger's four-way classification accuracy by 8.07 percentage points over a matched no-retrieval baseline that uses LLM-generated hints alone. Training ablations further support the combination of hard-example mining and Agent-enhanced rewards.
这样是否符合规则？每个自然段落用，长文用