refactor: improve submission instructions for clarity and user guidance in about.py

src/about.py

@@ -69,15 +69,20 @@ CITATION_BUTTON_TEXT = dedent("""
 SUBMISSION_INSTRUCTIONS_TEXT = dedent("""
 ## Submission Instructions
 
-0. **Login** to your HuggingFace account.
-1. Fill in the model name to search for on the HuggingFace Hub. (e.g. `qwen/qwen3-vl-8b-instruct`   )
-2. Select the model from search results, and check the model name autofilled below (e.g. `Qwen/Qwen3-VL-8B-Instruct`).
-3. (Optional) Fill in the revision commit of the model. If not filled, means using the latest `main` branch.
-4. Select the model type. (e.g. `pretrained`)
-5. Select the precision of the model. (e.g. `bfloat16`)
-6. Select the weights type of the model. (defaults to `Original`)
-7. (Optional) Fill in the base model name for **delta** or **adapter** weights. (e.g. `Qwen/Qwen3-VL-8B-Instruct`)
-8. Select a benchmark to evaluate on, and fill in the evaluation result value (e.g. `0.5` for `VSI-bench` `acc`).
-9. (Optional) click **[+]** button to add more benchmarks & evaluation result values.
-10. Click the **Submit Eval** button to submit the evaluation request.
+First, **Login** to your HuggingFace account, so that we can identify you and your evaluation results.
+
+Then, you can see the submission form, and fill in the following information:
+
+1. Fill in the **model name to search** for on the HuggingFace Hub. (e.g. `qwen/qwen3-vl-8b-instruct`   )
+2. Select the **model** from search results, and check the model name autofilled below (e.g. `Qwen/Qwen3-VL-8B-Instruct`).
+3. (Optional) Fill in the **revision commit** of the model. If not filled, means using the latest `main` branch.
+4. Select the **model type**. (e.g. `pretrained`)
+5. Select the **precision** of the model. (e.g. `bfloat16`)
+6. Select the **weights type** of the model. (defaults to `Original`)
+7. (Optional) Fill in the **base model name** for **delta** or **adapter** weights. (e.g. `Qwen/Qwen3-VL-8B-Instruct`)
+8. Check the checkbox of the **benchmark**s to evaluate on, and fill in the **evaluation result value** (e.g. `0.5` for `VSI-bench` `acc`).
+9. (Optional) Fill in the **commit message** to add extra message about the evaluation.
+10. Click the **[10. Submit Eval]** button to submit the evaluation request.
+
+Then your evaluation will be submitted, and our team will review it and add it to the leaderboard.
 """)