Installs as a private draft. Your edits and self-improvement runs do not change the published bundle.
Ensure reward functions produce consistent, well-scaled signals across different task types and difficulties
Ensure reward functions produce consistent, well-scaled signals across different task types and difficulties
Miscalibrated rewards cause training instability
Reward Engineering > Reward Function Design
--- name: "Reward Calibration" version: "1.0.0" description: "Ensure reward functions produce consistent, well-scaled signals across different task types and difficulties" tags: - rl-skills-universe - "reward-engineering" - "reward-function-design" trigger: "reward-calibration" --- # Reward Calibration Ensure reward functions produce consistent, well-scaled signals across different task types and difficulties ## Why It Matters Miscalibrated rewards cause training instability ## Domain **Reward Engineering** > Reward Function Design