Improve dataset card: Add links, sample usage, and metadata tags

README.md

@@ -1,18 +1,24 @@
 ---
-license: cc-by-4.0
-task_categories:
-- translation
 language:
 - zh
 - en
+license: cc-by-4.0
 size_categories:
 - n<1K
-
+task_categories:
+- translation
+tags:
+- machine-translation
+- cross-lingual
+- expert-domains
 ---
 
+# DiscoX Translation Benchmark
+
+[Paper](https://huggingface.co/papers/2511.10984) | [Project Page](https://randomtutu.github.io/DiscoX/) | [Code](https://github.com/ByteDance-Seed/DiscoX)
+
+DiscoX is a benchmark for the evaluation of LLMs on discourse- and expert-level translation tasks.
 
-# DiscoX Translation Benchmark 
-DiscoX is a benchmark for the evaluation of  LLMs on discourse- and expert-level translation tasks.
 ## Dataset At A Glance
 - **Languages**: English ⇄ Chinese (100 English→Chinese tasks, 100 Chinese→English tasks)
 - **Total samples**: 200 discourse- and exprt-level translation items
@@ -35,11 +41,35 @@ Secondary domain highlights include Social Scienices(社会科学),Natural Scien
   - `Primary_Domain`, `Secondary_Domain`: high-level topic labels
   - `prompt_id`, `__internal_uuid__`: identifiers for specific tasks
 
+## Sample Usage
+To get started with DiscoX, follow these steps:
+
+### 1. Install Dependencies
+Make sure you are using **Python 3.9+**. Then install the required packages:
+```bash
+pip install -r requirements.txt
+```
 
+### 2. Configure Environment Variables
+Set up your API key and endpoint in the `.env` file:
+```bash
+JUDGE_API_KEY=your_judgemodel_api_key_here
+JUDGE_API_BASE=your_judgemodel_api_base_here
+CANDIDATE_API_KEY=your_candidatemodel_api_key_here
+CANDIDATE_API_BASE=your_candidatemodel_api_base_here
+```
+
+### 3. Run Evaluation Tasks
+You can run tasks by specifying the target model and the judge model. For example:
+```bash
+python3 run_tasks.py --model openai/gpt4o-2024-11-20 --judgemodel azure/gemini-2.5-pro
+```
+
+### Example Use Case
+- **Model Under Evaluation:** `openai/gpt4o-2024-11-20`
+- **Judge Model:** `azure/gemini-2.5-pro`
+This configuration runs translation tasks using GPT-4o and evaluates them with Gemini-2.5-Pro under the Metric-S evaluation framework.
 
 ## Notes & Recommendations
 - The reference_list entries are designed to enable targeted verification of translation fidelity: by converting them into structured checks (e.g., terminology, tone, and named entities), the evaluation performs fine-grained, pointwise assessments of key translation aspects.
-- Translation instruction in pormpt describe desired output language in Chinese.
-
-## License
-Our data is under cc-by-4.0 license.
+- Translation instruction in pormpt describe desired output language in Chinese.