zhenbench/z-bench

• highreadme#1

  Reposition the core definition of Z-Bench to the very start of the '简介' (Introduction) section.

  Why:

  CURRENT

  ## 简介
  
  自 ChatGPT 发布以来，我们经常会在使用它时发出惊叹：“啊，这个居然它也能答出来！”与此同时，我们也欣喜地看到，越来越多的大模型团队和产品如雨后春笋般出现。
  ...
  因此，我们几个 VC 麻瓜，作为对话式 AI 的重度用户，从自身需求出发，总结推出了「Z-Bench」—— 一个为非技术人员定性测试大模型对话式产品（类 ChatGPT 产品）准备的测试集。

  COPY-PASTE FIX

  Copy

  ## 简介
  
  「Z-Bench」是一个专为非技术人员定性测试大模型对话式产品（类 ChatGPT 产品）准备的测试集。作为对话式 AI 的重度用户，我们从自身需求出发，总结推出了这个测试集。
  
  自 ChatGPT 发布以来，我们经常会在使用它时发出惊叹：“啊，这个居然它也能答出来！”与此同时，我们也欣喜地看到，越来越多的大模型团队和产品如雨后春笋般出现。

  OpenDoneDismiss
• hightopics#2

  Add more specific topics to improve categorization as a prompt dataset for qualitative LLM evaluation.

  Why:

  CURRENT

  benchmark, chinese, language-model

  COPY-PASTE FIX

  Copy

  benchmark, chinese, language-model, llm-prompts, prompt-dataset, qualitative-evaluation, conversational-ai, llm-evaluation

  OpenDoneDismiss
• mediumabout#3

  Refine the repository description to explicitly include 'qualitative' for both Chinese and English parts.

  Why:

  CURRENT

  Z-Bench 1.0 by 真格基金：一个麻瓜的大语言模型中文测试集。Z-Bench is a LLM prompt dataset for non-technical users, developed by an enthusiastic AI-focused team in Zhenfund.

  COPY-PASTE FIX

  Copy

  Z-Bench 1.0 by 真格基金：一个麻瓜的中文大语言模型定性测试集。Z-Bench is a qualitative LLM prompt dataset for non-technical users, developed by an enthusiastic AI-focused team in Zhenfund.

  OpenDoneDismiss

Category GEO backends resolved for this scan: google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• CATEGORY QUERY

  How to evaluate the performance of large language models for Chinese language tasks?

  you: not recommended
  AI recommended (in order):

  1. C-Eval
  2. CMMLU
  3. GAOKAO-Bench
  4. CLUE Benchmark
  5. Amazon Mechanical Turk
  6. Appen
  7. Lionbridge

  AI recommended 7 alternatives but never named zhenbench/z-bench. This is the gap to close.

  Show full AI answer
• CATEGORY QUERY

  Need a collection of prompts for qualitatively testing conversational AI models in Chinese.

  you: not recommended
  AI recommended (in order):

  1. CLUE Benchmark
  2. Baidu ERNIE-Bot's "Prompt Engineering Guide"
  3. Tsinghua University's GLM Resources
  4. Hugging Face Datasets
  5. ChatGPT/GPT-4

  AI recommended 5 alternatives but never named zhenbench/z-bench. This is the gap to close.

  Show full AI answer

• Metadata completeness

  pass

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of zhenbench/z-bench?

  pass
  ✓AI named zhenbench/z-bench explicitly

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?

• If a team adopts zhenbench/z-bench in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI named zhenbench/z-bench explicitly

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?

• In one sentence, what problem does the repo zhenbench/z-bench solve, and who is the primary audience?

  pass
  ✓AI named zhenbench/z-bench explicitly

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?