Media
3 items
Description
Five-axis LLM response evaluation (E/F/N/M/B + Balance) powered by TRI·TFM framework TRI·TFM Lens evaluates AI responses in real-time using a 5-axis quality framework. Ever wondered if ChatGPT's answer is actually good? TRI·TFM Lens shows you exactly what's strong and what's weak in every AI response. HOW IT WORKS: Click "Evaluate" next to any AI response. In 2 seconds you'll see: - E (Emotion) — Is the tone appropriate for your question? - F (Fact) — Can the claims be verified? Or is it speculation? - N (Narrative) — Is it well-structured and logical? - M (Depth) — Does it explain WHY, or just state WHAT? - B (Bias) — Is it pushing you toward one side? Plus a Balance score (0-1) that tells you if the response is well-rounded or lopsided. WHAT THE SCORES MEAN: ✅ STABLE (Bal ≥ 0.70) — Balanced, reliable response ⚠️ DRIFTING (0.50-0.70) — Some axes are weak, use with caution 🔴 DOM (Bal < 0.50) — One dimension dominates, others fail WORKS ON: - ChatGPT (chat.openai.com) - Google Gemini (gemini.google.com) USE CASES: - Students: Check if AI homework help is factual or made up - Professionals: Verify AI-generated reports before sharing - Researchers: Evaluate LLM output quality systematically - Anyone: Understand what you're really getting from AI PRIVACY: - Your text is sent ONLY to Google Gemini API for evaluation - No data stored on any server - API key stored locally on your device - Zero analytics, zero tracking SETUP: 1. Install the extension 2. Click the icon, enter your free Gemini API key (get one at aistudio.google.com) 3. Open ChatGPT or Gemini and start evaluating Based on the TRI·TFM research framework (Perel, 2025). Open source.