Update openevolve/evaluator.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Author: windyriki

openevolve/evaluator.py

@@ -208,7 +208,8 @@ async def evaluate_program(
                         if "combined_score" in eval_result.metrics:
                             # Original combined_score is just accuracy
                             accuracy = eval_result.metrics["combined_score"]
-                            # Combine with LLM average (70% accuracy, 30% LLM quality)
+                            # Combine accuracy with LLM average using dynamic weighting:
+                            # (1 - llm_feedback_weight) * accuracy + llm_feedback_weight * LLM quality
                             eval_result.metrics["combined_score"] = (
                                 accuracy * (1-self.config.llm_feedback_weight) + llm_average * self.config.llm_feedback_weight
                             )