Add and Enhance Chinese Translation of Techniques Section#729
Add and Enhance Chinese Translation of Techniques Section#729MochiaoChen wants to merge 9 commits intodair-ai:mainfrom
Conversation
Add meta-prompting entry to _meta.zh.json
Translate prompt_chaining and pal to Chinese
|
@MochiaoChen is attempting to deploy a commit to the DAIR-AI Team on Vercel. A member of the Team first needs to authorize it. |
There was a problem hiding this comment.
Pull request overview
This pull request enhances the Chinese localization of prompt engineering techniques documentation. It updates terminology for better clarity and consistency, refines the Active-Prompt section, improves citation formatting in the Reflexion section, and adds a comprehensive new section on Meta Prompting (元提示词).
Changes:
- Updated Chinese translations for several technique names in the navigation menu, replacing English terms with clearer Chinese equivalents
- Revised the Active-Prompt page to use consistent Chinese terminology "主动提示" throughout
- Added a complete new documentation page for Meta Prompting with detailed explanations of its characteristics, advantages, and applications
- Updated citation formatting in the Reflexion page
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.
| File | Description |
|---|---|
| pages/techniques/_meta.zh.json | Updates navigation menu with improved Chinese translations for technique names including "链式提示", "主动提示", "PAL（程序辅助语言模型）", "自我反思（Reflexion）", and adds new entry "元提示词（Meta Prompting）" |
| pages/techniques/activeprompt.zh.mdx | Changes title from "Active-Prompt" to "主动提示（Active-Prompt）" and updates content to use consistent Chinese terminology |
| pages/techniques/reflexion.zh.mdx | Updates citation format from [Shinn et al. (2023)] to [Shinn等人， (2023)] for Chinese localization |
| pages/techniques/meta-prompting.zh.mdx | Adds comprehensive new documentation page covering Meta Prompting concept, characteristics, advantages over few-shot prompting, and application scenarios with course promotion components |
Comments suppressed due to low confidence (1)
pages/techniques/activeprompt.zh.mdx:12
- Inconsistent spacing in citation format within the same file. Line 7 uses
[Diao 等人（2023）](with space before 等人), while line 12 uses[Diao等人（2023）](without space). For consistency within the file, both citations should use the same spacing pattern.
思维链（CoT）方法依赖于一组固定的人工注释范例。问题在于，这些范例可能不是不同任务的最有效示例。为了解决这个问题，[Diao 等人（2023）](https://arxiv.org/pdf/2302.12246.pdf)最近提出了一种新的提示方法，称为主动提示，以适应 LLMs 到不同的任务特定示例提示（用人类设计的 CoT 推理进行注释）。
下面是该方法的说明。第一步是使用或不使用少量 CoT 示例查询 LLM。对一组训练问题生成 *k* 个可能的答案。基于 *k* 个答案计算不确定度度量（使用不一致性）。选择最不确定的问题由人类进行注释。然后使用新的注释范例来推断每个问题。
<Screenshot src={ACTIVE} alt="ACTIVE" />
图片来源：[Diao等人（2023）](https://arxiv.org/pdf/2302.12246.pdf)
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| 根据[Zhang等人，2024](https://arxiv.org/abs/2311.11482)的研究，元提示词的核心特性可归纳如下： | ||
|
|
||
| **1. 结构导向**: 优先考虑问题和解决方案的格式与模式，而非具体内容。 | ||
|
|
||
| **2. 句法聚焦**: 利用句法作为预期回答或解决方案的引导模板。 | ||
|
|
||
| **3. 抽象示例**: 使用抽象示例作为框架，展示问题和解决方案的结构，而不关注具体细节。 | ||
|
|
||
| **4. 多用途**: 适用于各种领域，能够针对广泛的问题提供结构化的响应。 | ||
|
|
||
| **5. 分类方法**: 借助类型理论，强调提示中组件的分类和逻辑排列。 | ||
|
|
||
| ## 与少样本提示 (Few-Shot Prompting) 的优势对比 | ||
|
|
||
| [Zhang等人, 2024](https://arxiv.org/abs/2311.11482) 指出，元提示词与少样本提示的不同之处在于，元提示词侧重于结构导向的方法，而少样本提示则强调内容驱动的方法。 | ||
|
|
||
| 以下示例(原自[Zhang等人， 2024](https://arxiv.org/abs/2311.11482) )展示了在解决 MATH 基准测试问题时，结构化元提示词与少样本提示词之间的区别： |
There was a problem hiding this comment.
Citation format is inconsistent within this file and with the rest of the codebase. Lines 10, 24, and 26 use different formats: [Zhang等人，2024] (no space after comma), [Zhang等人, 2024] (space after comma), and [Zhang等人， 2024] (full-width comma with space). The established pattern in this codebase is to use Chinese parentheses around the year without a comma: [Author等人（YEAR）] (e.g., [Diao 等人（2023）] in activeprompt.zh.mdx:7, [Wang等人（2022）] in consistency.zh.mdx:3). All three citations should use the format [Zhang 等人（2024）] for consistency.
| "multimodalcot": "多模态思维链提示方法", | ||
| "graph": "基于图的提示" | ||
| "graph": "基于图的提示", | ||
| "meta-prompting": "元提示词（Meta Prompting）" |
There was a problem hiding this comment.
The ordering of "meta-prompting" is inconsistent with the English version. In the English _meta.en.json, "meta-prompting" appears on line 5 (after "cot" and before "consistency"), but in the Chinese version it's placed at line 19 (at the end of the list). This ordering difference may affect the navigation menu structure and user experience. Consider aligning the order with the English version by placing "meta-prompting" between "cot" and "consistency" for consistency across language versions.
| # 自我反思（Reflexion） | ||
|
|
||
| 自我反思是一个通过语言反馈来强化基于语言的智能体的框架。根据 [Shinn et al. (2023)](https://arxiv.org/pdf/2303.11366.pdf)，“自我反思是一种‘口头’强化的新范例，它将策略参数化为智能体的记忆编码与 LLM 的参数选择配对。” | ||
| 自我反思是一个通过语言反馈来强化基于语言的智能体的框架。根据 [Shinn等人， (2023)](https://arxiv.org/pdf/2303.11366.pdf)，“自我反思是一种‘口头’强化的新范例，它将策略参数化为智能体的记忆编码与 LLM 的参数选择配对。” |
There was a problem hiding this comment.
Citation format is inconsistent with the rest of the codebase. The current format [Shinn等人， (2023)] uses a comma followed by a space before the parentheses. However, the established pattern in this codebase is to use [Author等人（YEAR）] with Chinese parentheses and no comma (e.g., [Diao 等人（2023）] in activeprompt.zh.mdx:7, [Wang等人（2022）] in consistency.zh.mdx:3, [Wei等人（2022）] in cot.zh.mdx:13,15). The format should be [Shinn 等人（2023）] to match the established convention.
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
This pull request introduces improvements to the Chinese localization and documentation of prompt engineering techniques. The main updates include refining terminology for several techniques, updating the content and naming for "主动提示" (Active-Prompt), and adding a comprehensive new section for "元提示词" (Meta Prompting). These changes enhance clarity, consistency, and educational value for users learning about advanced prompting methods.
Terminology and Translation Improvements:
pages/techniques/_meta.zh.jsonto provide clearer and more consistent Chinese terminology, including "链式提示" for Prompt Chaining, "主动提示" for Active-Prompt, and "PAL（程序辅助语言模型）" for PAL. Added a new entry for "元提示词（Meta Prompting）".Content and Documentation Enhancements:
pages/techniques/activeprompt.zh.mdxto use "主动提示" consistently and clarify the explanation of the method.pages/techniques/meta-prompting.zh.mdxdescribing the concept, features, advantages, and application scenarios of "元提示词（Meta Prompting）", including visual examples and links to related courses.pages/techniques/reflexion.zh.mdxby updating author references and maintaining consistent terminology.