Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Sign in
Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

A page, a video or an image online.

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

GRPO and RLVR training on verifiable rewards v1.0.0

Use this resource to train a model on verifiable-reward tasks with GRPO

New

  • 548 downloads
  • 125 likes
  • 203 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install grpo-verifiable-reward-training

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent first confirms RL is the right tool, writes the GRPO configuration with a verifier (code execution, tests, schema) and inspects rewards before launch. It also helps choose a variant if the run diverges or reward-hacks.

Mostly picked by

Updated November 2, 2026

Card for the “GRPO and RLVR training on verifiable rewards” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
MIT

Similar resources

View all

What you get

  • SKILL.md: recipe, inspection rule and variant selection
  • references/reward-functions.md: reward functions
  • references/grpo-memory.md: GRPO memory management

Good fit for

  • Your task can be checked by a test, a schema or an execution
  • Your GRPO run diverges or games its reward

Not for

    Tags

    • grpo
    • rlvr
    • reward
    • reinforcement learning

    About

    You say

    “Set up GRPO where the reward is passing the generated unit tests.”

    Card for the “GRPO and RLVR training on verifiable rewards” skill

    Skill published by wshobson under the MIT license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to train a model on verifiable-reward tasks with GRPO
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/grpo-memory.md3.6 KB
    • references/reward-functions.md11.8 KB
    • SKILL.md7.8 KB

    Origin and licence

    Listed unmodified

    Author
    wshobson
    Source
    github.com/wshobson/agents/tree/0874a27b17a22e2b7c2e2bcd725768e0803cad06/plugins/llm-finetuning/skills/grpo-rlvr-training
    Licence
    MIT

    Notice

    « grpo-rlvr-training » (wshobson, MIT) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.