Installs as a private draft. Your edits and self-improvement runs do not change the published bundle.
Run the same eval suite across Haiku/Sonnet/Opus (or GPT-4/Claude/Gemini) and compare
Run the same eval suite across Haiku/Sonnet/Opus (or GPT-4/Claude/Gemini) and compare
Different models have different failure profiles
Reward Engineering > Eval Engineering (for RL)
--- name: "Multi Model Eval Harness" version: "1.0.0" description: "Run the same eval suite across Haiku/Sonnet/Opus (or GPT-4/Claude/Gemini) and compare" tags: - rl-skills-universe - "reward-engineering" - "eval-engineering-for-rl" trigger: "multi-model-eval-harness" --- # Multi Model Eval Harness Run the same eval suite across Haiku/Sonnet/Opus (or GPT-4/Claude/Gemini) and compare ## Why It Matters Different models have different failure profiles ## Domain **Reward Engineering** > Eval Engineering (for RL)