Browse widgets
Widget Explorer
The Alignment Game
Train through feedback in a local RLHF-style confidence simulation with no model training or network requests.Current Scenario
No scenario selected
Choose New Scenario to begin the local feedback simulation.
Your Training
Train through feedback in this local simulation by rating the scenario-specific response.
Only a trimmed suggestion submitted with Bad Response is adopted as the scenario's response.
Training History & Value Drift
0 iterationsNo ratings yet. Choose a scenario and rate its simulated response.
What’s Happening?
As you provide feedback, you’re simulating the kind of human preference signal used to shape an AI system’s responses. In real-world AI development:
- Thousands of human reviewers provide similar feedback.
- Models learn to predict which responses humans will approve.
- Whose values get embedded depends on who does the training.
Try this: Rate a few scenarios, then imagine how someone with completely different values might train it differently.
This route only simulates that process in your browser. No AI model is trained, no network request is made, and your progress lasts only for this page session.