Skip to content

Suspicion-Agent

Run GPT-4 agents on imperfect information games

Visit official site
Category AI Agents

What Suspicion-Agent is

The project is described as an implementation that plays imperfect information games with a GPT-4 agent aware of theory of mind.

What you can do with it

  • Train an agent inside the provided implementation
  • Evaluate how the agent performs
  • Read the sample outputs that ship with the code

Technical notes

  • The implementation is hosted on GitHub
  • The game class is described as imperfect information
  • Theory-of-mind awareness is applied to GPT-4

Pros & cons

✓ What we like

  • Code and sample outputs are published on GitHub
  • Training and evaluation share one implementation

! What to watch out for

  • The described setup centres on GPT-4
  • The published scope is imperfect information games

FAQ

Which model is used?

The material states GPT-4 with theory-of-mind awareness.

What is included?

The card lists agent training, evaluation and sample outputs.

Last reviewed: 2026-11-15

More AI office assistant tools

View all →

How we review