Individual skill

grpo

Reference for the GRPO (Group Relative Policy Optimization) algorithm. Use when implementing, debugging, or verifying a GRPO training pipeline — covers the mathematical formulation (group-relative advantages, clipped surrogate loss, KL penalty), the training loop (generate → score → advantage → loss), log-probability computation, advantage estimation, and relationship to PPO/REINFORCE.

tasks/debug-trl-grpo/environment/skills/grpo/SKILL.md

Read pinned source

Source commit f860ecf02ef53412a030f8661c83d644b739ff01

What was checked

Matches the observed repository HEAD: Yes
Scanner rules are current: Yes
Repository scan complete: No
Skill scan complete: Yes
Observed behavior: Instructions only
Last checked: Nov 9, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings): 1,911
Repository last updated: Nov 4, 2026
License: Apache-2.0

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

benchmark Repository topic
skillsbench Repository topic
testing Derived category

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files

SKILL.mdcore · 4,441 bytes
references/grpo-algorithm.mddoc · 4,980 bytes
references/grpo-trainer-internals.mddoc · 3,954 bytes

Read excerpt


# GRPO Algorithm Reference

## Overview

Group Relative Policy Optimization (GRPO) is a policy gradient method that eliminates the need for a learned critic by computing advantages from group statistics. For each prompt, multiple completions are sampled and their rewards are normalized within the group to produce advantages.

**Key insight:** Instead of training a value function to estimate `V(s)`, GRPO uses the mean reward of the group as the baseline. This removes the critic entirely, reducing memory and avoiding value function approximation errors.

## Training Loop

Each step follows this pipeline:

```
1. Sample G completions per prompt from current policy