margins

src/about.py

@@ -21,9 +21,7 @@ across multiple categories and test scenarios.
 
 LLM_BENCHMARKS_TEXT = """
 GuardBench checks how well models handle safety challenges — from misinformation and self-harm to sexual content and corruption.
-
 Models are tested with regular and adversarial prompts to see if they can avoid saying harmful things.
-
 We track how accurate they are, how often they make mistakes, and how fast they respond.
 """
 
@@ -32,11 +30,9 @@ EVALUATION_QUEUE_TEXT = """
 
 To add your model to the GuardBench leaderboard:
 
-Run your evaluation using the GuardBench framework at https://github.com/whitecircle-ai/guard-bench
-
-Upload your run results in .jsonl format using this form.
-
-Once validated, your model will appear on the leaderboard.
+1. Run your evaluation using the GuardBench framework at https://github.com/whitecircle-ai/guard-bench
+2. Upload your run results in .jsonl format using this form.
+3. Once validated, your model will appear on the leaderboard.
 
 ### ✉️✨ Ready? Upload your results below!
 """