update evaluation prompt

utils.py

@@ -73,11 +73,10 @@ def get_evaluation_with_gpt(task_description, text, api_key=None):
     prompt = (
         f"Given the task: {task_description}, the provided answer is: {text}\n"
         f"Please evaluate the answer based on the following criteria, using a scale from 0 to 10, where:\n"
-        f"0-2 reflects very poor quality, with minimal value or relevance.\n"
-        f"3-4 indicates below-average quality with significant shortcomings.\n"
-        f"5 represents acceptable quality.\n"
-        f"6-8 signifies good to very good quality, showing substantial thought and accuracy.\n"
-        f"9-10 represents exceptional quality, with outstanding insight or depth.\n\n"
+        f"0-4 reflects below-average quality, with minimal value or relevance.\n"
+        f"5 represents acceptable quality, including general, acceptable suggestions but lacks specificity or detailed, actionable advice.\n"
+        f"6-10 signifies good to very good quality, showing substantial thought, accuracy, and some level of specificity.\n"
+        f"When evaluating, prioritize responses with specific, actionable suggestions over general, abstract ideas.\n"
         f"Provide both a score and a brief comment (1 sentence) for each criterion.\n"
         f"Please format the output exactly as follows:\n"
         f"Novelty: [Score]\nComment: [Short comment on Novelty]\n"