Add paper link and metadata to QuestBench

This PR improves the dataset card for QuestBench by:
- Adding YAML metadata for `task_categories` and `language`.
- Linking the dataset to the associated research paper: [Teaching AI Through Benchmark Construction: QuestBench as a Course-Based Practice for Accountable Knowledge Work](https://huggingface.co/papers/2605.21413).
- Maintaining the existing file descriptions and execution instructions.

README.md

@@ -1,12 +1,15 @@
+---
+task_categories:
+- text-generation
+language:
+- zh
+---
+
 # QuestBench
 
-**QuestBench** is a course-curated benchmark of **256 expert-level questions**
-for evaluating the deep-search capabilities of language models. Questions are
-long-tail queries spanning **14 normalized domains** in the humanities and
-social sciences (Literature, Law, History, Foreign Languages, Social
-Sciences, Arts, Archaeology, Journalism, …). Each question, reference answer,
-and grading rubric is in Chinese; the model under evaluation may search and
-read in any language.
+**QuestBench** is a course-curated benchmark of **256 expert-level questions** for evaluating the deep-search capabilities of language models, introduced in the paper [Teaching AI Through Benchmark Construction: QuestBench as a Course-Based Practice for Accountable Knowledge Work](https://huggingface.co/papers/2605.21413). 
+
+Questions are long-tail queries spanning **14 normalized domains** in the humanities and social sciences (Literature, Law, History, Foreign Languages, Social Sciences, Arts, Archaeology, Journalism, …). Each question, reference answer, and grading rubric is in Chinese; the model under evaluation may search and read in any language.
 
 ## Files
 
@@ -15,8 +18,7 @@ read in any language.
 - `vision_server.py` — local MCP server proxying image analysis to OpenRouter.
 - `package.json`, `package-lock.json` — npm dependencies.
 
-Each JSONL record has four fields: `domain`, `question`, `answer`,
-`grading_criteria` (the rubric an LLM judge uses to assign a 0–100 score).
+Each JSONL record has four fields: `domain`, `question`, `answer`, `grading_criteria` (the rubric an LLM judge uses to assign a 0–100 score).
 
 ## Run
 
@@ -40,4 +42,4 @@ QUESTION=$(jq -r '.question' questbench.jsonl | head -n 1) && \
     --permission-mode bypassPermissions \
     --disallowedTools WebSearch WebFetch \
     --output-format stream-json --verbose
-```
+```