Knapsack Rl Budget Allocation

Improve RL training for LLMs by dynamically allocating exploration budget (rollout count) to tasks based on their difficulty and current learning status. Solves the knapsack problem of maximizing gradient signal within fixed compute budget, increasing non-zero policy gradients by 20-40% and achieving 2-4 point performance gains.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit cba7f95505

Frequently asked questions

npx get-skill add adu2021/knapsack-rl-budget-allocation