Update src/about.py

src/about.py

@@ -22,11 +22,10 @@ accurate, and actionable across multiple programming languages and review catego
 LLM_BENCHMARKS_TEXT = """
 CodeReview Bench is a comprehensive benchmark for evaluating automated code review systems across programming languages and review quality dimensions.
 
-It evaluates models on their ability to provide high-quality code reviews using both LLM-based multimetric evaluation (readability, relevance, explanation clarity, problem identification, actionability, completeness, specificity, contextual adequacy, consistency, brevity) and exact-match metrics (pass@1, pass@5, pass@10, BLEU@10).
+It evaluates models on their ability to provide high-quality code reviews using both LLM-based multimetric evaluation (readability, relevance, explanation clarity, problem identification, actionability, completeness, specificity, contextual adequacy, consistency, brevity) and exact-match metrics (pass@1, pass@5, pass@10) presented in our paper.
 
-The benchmark supports both Russian and English comment languages across 17+ programming languages including Python, JavaScript, Java, C++, TypeScript, Go, Rust, and more.
+The benchmark supports both Russian and English comment languages across 4 programming languages including Python, Java, Go, Scala
 
-Learn more about automated code review evaluation and best practices.
 """
 
 EVALUATION_QUEUE_TEXT = """