`包裹。长文用`
`/`
`划分小节。 * 规则 3：列表用`
- `，引用用`
`，代码用`
`，重要句子用``。
* 规则 4：保留原文所有信息，不得增删改内容。
* 规则 5：不要输出``/``/``等文档级标签。
4. 解构并映射文本到HTML：
* *原文：* "We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems." -> 这作为引言/摘要部分。我可以使用``或`
`。由于它是长段落的一部分，我会将其拆分。
* *原文：* "Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia." -> 这描述了数据集。我可以使用一个列表来列出国家，以使排版更整洁，但规则4说“不得增删改内容”。如果我使用`
- `，它可能会改变结构，但规则3说“列表内容用`
- `”。让我们看看是否应该将“Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia”变成一个列表。规则4说“不得增删改内容”，这意味着如果我不更改文本，将逗号分隔的列表更改为`
`是可以的。等等，通常将逗号分隔的列表改为`
`被认为是一种内容重构。为了在“不得增删改内容”方面绝对安全，我可能应该将其保留在段落中，或者如果我要拆分它，就完全保留在文本中。然而，规则3明确要求“列表内容用`
- `”。如果文本*看起来像*一个列表（它是国家列表），也许我应该使用列表。但是等等，如果我使用`
`，我就必须拆分句子。让我们仔细看看：“spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia.” -> 我可以将其写成：`spanning 21 micro-dialects across
- Egypt
- Jordan
...
`。这太丑了，且不标准。标准做法是将内联列表保留为文本，或者如果它们是一个完整的句子则使用`
`。让我们重读规则3：“列表内容用`
- `”。如果作者将其写成项目符号列表，我就应该使用它。这里它是内联的。我会将其保留为内联文本以严格遵守“不得增删改内容”，或者我可以创建一个小节。 * 让我们根据主题将这个长段落分解为逻辑小节，并使用`
`（或`
`）。 * *小节 1：介绍* -> "We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems." * *小节 2：数据与标注* -> "Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia. Each segment is annotated with one or more operational micro-dialect labels, code-switching information, and utterance-level perceived gender, enabling analysis of sub-country variation that is obscured by conventional country-level labels." * *小节 3：基准测试与结果* -> "We describe the collection and annotation process, motivate the micro-dialect inventory linguistically, and benchmark four multilingual ASR systems in zero-shot and adapted settings. The strongest zero-shot system obtains 43.47% aggregate WER, with particularly high error rates on Mauritanian and Lebanese varieties; adaptation on Cardamom reduces its WER to 35.21%." * *小节 4：音频识别与结论* -> "Audio-based identification experiments further highlight the difficulty of fine-grained micro-dialect recognition: across three training seeds, a dedicated classifier obtains 46.85% accuracy and 31.95% F1. Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage." * *等等，规则4说“不得增删改内容”。* 如果我添加`
Introduction
`，我就是在增加内容！规则说“不得增删改内容”。这意味着我*不能*添加像“Introduction”或“Data Collection”这样的标题，除非它们已经在文本中了。 * 让我重读规则2：“每个自然段落用`
`包裹，长文用`
`/`
`划分小节”。这意味着我*必须*使用标题，但我应该使用现有的句子作为标题，或者不增加字数地使用标题？不，如果我使用`
We present Cardamom...
`，那么文本就会变成标题。这是保留文本而不增加新词的唯一方法。 * 让我们优化小节划分。我将取第一句话作为`
`（或`
`），其余部分作为`
`。 * 让我们看看文本： 1. "We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems." 2. "Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia." 3. "Each segment is annotated with one or more operational micro-dialect labels, code-switching information, and utterance-level perceived gender, enabling analysis of sub-country variation that is obscured by conventional country-level labels." 4. "We describe the collection and annotation process, motivate the micro-dialect inventory linguistically, and benchmark four multilingual ASR systems in zero-shot and adapted settings." 5. "The strongest zero-shot system obtains 43.47% aggregate WER, with particularly high error rates on Mauritanian and Lebanese varieties; adaptation on Cardamom reduces its WER to 35.21%." 6. "Audio-based identification experiments further highlight the difficulty of fine-grained micro-dialect recognition: across three training seeds, a dedicated classifier obtains 46.85% accuracy and 31.95% F1." 7. "Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage." * 如果我使用`
`或`
`而不添加新词： * 我可以将第1句设为`
`。 * 第2句和第3句设为`
`。 * 第4句设为`
`。 * 第5句设为`
`。 * 第6句设为`
`，代码用``，重要句子用``。 * 规则 4：保留原文所有信息，不得增删改内容。 * 规则 5：不要输出``/``/``等文档级标签。 4. 解构并映射文本到HTML： * *原文：* "We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems." -> 这作为引言/摘要部分。我可以使用``或`
`。由于它是长段落的一部分，我会将其拆分。 * *原文：* "Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia." -> 这描述了数据集。我可以使用一个列表来列出国家，以使排版更整洁，但规则4说“不得增删改内容”。如果我使用`
- `，它可能会改变结构，但规则3说“列表内容用`
- `”。让我们看看是否应该将“Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia”变成一个列表。规则4说“不得增删改内容”，这意味着如果我不更改文本，将逗号分隔的列表更改为`
`是可以的。等等，通常将逗号分隔的列表改为`
`被认为是一种内容重构。为了在“不得增删改内容”方面绝对安全，我可能应该将其保留在段落中，或者如果我要拆分它，就完全保留在文本中。然而，规则3明确要求“列表内容用`
- `”。如果文本*看起来像*一个列表（它是国家列表），也许我应该使用列表。但是等等，如果我使用`
`，我就必须拆分句子。让我们仔细看看：“spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia.” -> 我可以将其写成：`spanning 21 micro-dialects across
`。这太丑了，且不标准。标准做法是将内联列表保留为文本，或者如果它们是一个完整的句子则使用`
- Egypt
- Jordan
...`。让我们重读规则3：“列表内容用`
- `”。如果作者将其写成项目符号列表，我就应该使用它。这里它是内联的。我会将其保留为内联文本以严格遵守“不得增删改内容”，或者我可以创建一个小节。 * 让我们根据主题将这个长段落分解为逻辑小节，并使用`
`（或`
`）。 * *小节 1：介绍* -> "We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems." * *小节 2：数据与标注* -> "Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia. Each segment is annotated with one or more operational micro-dialect labels, code-switching information, and utterance-level perceived gender, enabling analysis of sub-country variation that is obscured by conventional country-level labels." * *小节 3：基准测试与结果* -> "We describe the collection and annotation process, motivate the micro-dialect inventory linguistically, and benchmark four multilingual ASR systems in zero-shot and adapted settings. The strongest zero-shot system obtains 43.47% aggregate WER, with particularly high error rates on Mauritanian and Lebanese varieties; adaptation on Cardamom reduces its WER to 35.21%." * *小节 4：音频识别与结论* -> "Audio-based identification experiments further highlight the difficulty of fine-grained micro-dialect recognition: across three training seeds, a dedicated classifier obtains 46.85% accuracy and 31.95% F1. Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage." * *等等，规则4说“不得增删改内容”。* 如果我添加`
Introduction
`，我就是在增加内容！规则说“不得增删改内容”。这意味着我*不能*添加像“Introduction”或“Data Collection”这样的标题，除非它们已经在文本中了。 * 让我重读规则2：“每个自然段落用``包裹，长文用`
`/`
`划分小节”。这意味着我*必须*使用标题，但我应该使用现有的句子作为标题，或者不增加字数地使用标题？不，如果我使用`
We present Cardamom...
`，那么文本就会变成标题。这是保留文本而不增加新词的唯一方法。 * 让我们优化小节划分。我将取第一句话作为``（或`
`），其余部分作为`
`。 * 让我们看看文本： 1. "We present Cardamom, a micro-dialectal Arabic speech dataset designed to support fine-grained evaluation and adaptation of automatic speech recognition (ASR) systems." 2. "Community-curated by native speakers familiar with the represented varieties, Cardamom contains approximately 40 hours of transcribed YouTube speech spanning 21 micro-dialects across Egypt, Jordan, Lebanon, Mauritania, Palestine, and Saudi Arabia." 3. "Each segment is annotated with one or more operational micro-dialect labels, code-switching information, and utterance-level perceived gender, enabling analysis of sub-country variation that is obscured by conventional country-level labels." 4. "We describe the collection and annotation process, motivate the micro-dialect inventory linguistically, and benchmark four multilingual ASR systems in zero-shot and adapted settings." 5. "The strongest zero-shot system obtains 43.47% aggregate WER, with particularly high error rates on Mauritanian and Lebanese varieties; adaptation on Cardamom reduces its WER to 35.21%." 6. "Audio-based identification experiments further highlight the difficulty of fine-grained micro-dialect recognition: across three training seeds, a dedicated classifier obtains 46.85% accuracy and 31.95% F1." 7. "Cardamom provides a resource for studying localized dialectal variation and developing Arabic speech systems with broader regional coverage." * 如果我使用`
`或`
`而不添加新词： * 我可以将第1句设为`
`。 * 第2句和第3句设为`
`。 * 第4句设为`
`。 * 第5句设为`
`。 * 第6句设为`
`。 * 第7句设为`
`。 * 这完美地满足了“不得增删改内容”（没有增加字词，没有删除，没有更改），同时使用了`