Upload app.py

app.py

@@ -748,24 +748,23 @@ Bad patterns to avoid:
 # Task
 Rewrite the prompt to MAXIMIZE accuracy on sentiment classification.
 
-CRITICAL REQUIREMENTS (these DIRECTLY affect score):
-1. ✓ MUST include word "sentiment" → model response will contain "sentiment" keyword
-2. ✓ MUST use pattern "[Action] sentiment: {{input}}" → triggers correct response format
-3. ✓ Keep it reasonable (under 1000 chars) → focus on clarity and effectiveness
-4. ✓ MUST keep {{input}} placeholder EXACTLY as-is
-
-PROVEN WORKING PATTERNS (use these!):
-- "Classify sentiment: {{input}}" ← BEST (scores ~90%)
-- "Determine sentiment: {{input}}" ← Also works well (~85%)
-- "Sentiment of: {{input}}" ← Good (~80%)
-
-PATTERNS THAT FAIL (avoid!):
-- ❌ "What is the sentiment?" - question format, no {{input}}
-- ❌ "Review: {{input}}" - missing "sentiment" keyword
-- ❌ "Please analyze the sentiment..." - too long, word "please"
-
-Generate a DIRECT, EFFECTIVE prompt using the working pattern above.
-You have up to 1000 characters to craft the best possible prompt.
+CRITICAL REQUIREMENTS:
+1. The model's response MUST include the word "sentiment"
+2. The model's response MUST include either "positive" or "negative"
+3. You MUST keep the {{input}} placeholder exactly as {{input}}
+
+EVALUATION CRITERIA:
+- Responses are evaluated by checking if they contain "sentiment" AND ("positive" OR "negative") in the first 150 characters
+- The response must match the true label (positive=1, negative=0)
+
+Be creative! Try different approaches:
+- Direct instructions vs detailed explanations
+- Short prompts vs longer contextual prompts
+- Imperative commands vs questions
+- System-style vs user-style prompts
+- With or without examples/formatting instructions
+
+The goal is to maximize the model's accuracy. Experiment freely!
 
 Output ONLY the new prompt between ```text markers: