📅 2025-06-18 — Session: Onyx Hammer Project and AI Prompt Engineering
🕒 23:30–23:55
🏷️ Labels: AI, Prompt Engineering, Onyx Hammer, Physics, Rubric Design
📂 Project: Dev
⭐ Priority: MEDIUM
Session Goal
The session focused on two primary objectives: providing a comprehensive project overview for Onyx Hammer and developing a robust strategy for AI prompt engineering and evaluation.
Key Activities
- Project Overview for Onyx Hammer: Detailed the core mission, compensation model, current status, and recommendations for proceeding with the Onyx Hammer project.
- Prompt Engineering Strategy: Outlined a strategy for crafting expert-level prompts and evaluation rubrics to assess AI robustness, particularly in the Physics domain.
- Crafting AI Prompts: Developed complex AI prompts for modeling atmospheric entry dynamics, focusing on expert reasoning and tool use.
- Quiz Analysis for Onyx Hammer: Analyzed quiz items from Step 3 of the Onyx Hammer framework, providing correct responses and rationale.
- Rubric Construction: Created effective rubric criteria for evaluating AI responses, ensuring clarity and alignment with prompt demands.
Achievements
- Established a clear project plan and evaluation strategy for Onyx Hammer.
- Developed expert-level AI prompts and rubrics for assessing AI models in Physics.
Pending Tasks
- Implement the recommendations and strategies outlined for Onyx Hammer.
- Conduct further testing and refinement of AI prompts and rubrics.