PegJump

Description

PegJump is an environment for evaluating agents on strategic planning and sequential reasoning. This environment wraps the PegJump implementation from TextArena, a framework for text-based game environments.

Capabilities

• Strategic long-term planning
• Sequential move optimization
• Spatial reasoning on board configurations
• Backtracking and alternative path exploration

Compute Requirements

PegJump does not require a sandbox. It has minimal compute requirements.

License

MIT.

Tasks

There are two splits: train (150 tasks) and test (150 tasks). Each split contains 50 tasks across each of 3 variants:

• PegJump-v0
• PegJump-v0-train
• PegJump-v0-raw

Each task is seeded for reproducibility.

Reward Structure

This is a sparse reward environment. Rewards are mapped from TextArena's native range of {-1, 0, 1} to {0.0, 0.5, 1.0} via (raw + 1) / 2.

We do not use LLM graders for this environment; reward is determined programmatically.

Data

Game state is generated procedurally by the TextArena engine using seeded randomness. No external data files are required.

Tools

Agents are given a single tool:

• jump(from_hole, to_hole): Jump a peg from one hole to another. The peg jumps over an adjacent peg into an empty hole, removing the jumped peg.

Time Horizon

PegJump is a multi-turn environment.

Environment Difficulty

This environment presents moderate to challenging difficulty, requiring agents to plan sequences of moves that lead to optimal final configurations with minimal remaining pegs.

Other Environment Requirements

There are no further environment requirements; PegJump works out of the box without any secrets or API keys.

Safety

Agents in PegJump interact only with a puzzle game and have no access to external systems, the internet, or sensitive data. The environment does not present safety risks.

Citations

@software{textarena2024,
  author    = {Guertler, Leon and Banting, Wilfried and Pignatelli, Eduardo},
  title     = {TextArena},
  year      = {2024},
  publisher = {GitHub},
  url       = {https://github.com/LeonGuertler/TextArena}
}