Skip to main content
Portfolio
The Illustrated RLHF Guide

RLHF Illustrated Guide

Learn Reinforcement Learning from Human Feedback through interactive visualizations, intuitive analogies, and hands-on examples.

Start LearningTry Playground

Analogy Visuals

These illustrations introduce the narrative lenses we use across the guide—browse them to understand each storytelling perspective.

Retro policy training

Atari Game Bot

Treat RLHF like a classic arcade challenge – policies learn by chasing new highs.

Editing draft…
Draft, edit, refine

Creative Writing Student

Follow the writing student and mentor through iterative feedback and revisions.

∑ reward_t ⋅ γ^tProof Verified
Whiteboard your steps

Math Tutor Bot

Trace every deduction on a collaborative whiteboard to keep reasoning grounded.

Push the frontier

Advanced Concepts

Peek at frontier alignment topics like Constitutional AI and tool-use orchestration.

Learning Path

Introduction to RLHF

What RLHF does and why it matters

Start here

Reward Modeling

Training reward models from preference data

View module

Policy Gradients (PPO)

Core RL optimization techniques

View module