# Puzzle Solving without Search or Human Knowledge: An Unnatural Language Approach

arXiv: https://arxiv.org/abs/2109.02797 · HTML: https://mlsanity.com/paper/2109.02797/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 33 references that are in the map.
- First posted on arXiv 7 Sep 2021.

## Builds on (most cited first, 8 of 8)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Generating Long Sequences with Sparse Transformers](https://mlsanity.com/paper/1904.10509/) (2019)
- [A Multiscale Visualization of Attention in the Transformer Model](https://mlsanity.com/paper/1906.05714/) (2019)
- [Understanding the Difficulty of Training Transformers](https://mlsanity.com/paper/2004.08249/) (2020)
- [The Chess Transformer: Mastering Play using Generative Language Models](https://mlsanity.com/paper/2008.04057/) (2020)
- [Solving the Rubik's Cube Without Human Knowledge](https://mlsanity.com/paper/1805.07470/) (2018)
- [The Go Transformer: Natural Language Modeling for Game Play](https://mlsanity.com/paper/2007.03500/) (2020)
- [Solving the Rubik's Cube Optimally is NP-complete](https://mlsanity.com/paper/1706.06708/) (2017)

## Built on by (most cited first, 8 of 8)

- [Puzzle Solving using Reasoning of Large Language Models: A Survey](https://mlsanity.com/paper/2402.11291/) (2024)
- [Causal Language Modeling Can Elicit Search and Reasoning Capabilities on Logic Puzzles](https://mlsanity.com/paper/2409.10502/) (2024)
- [CayleyPy RL: Pathfinding and Reinforcement Learning on Cayley Graphs](https://mlsanity.com/paper/2502.18663/) (2025)
- [PuzzlePlex: Benchmarking Foundation Models on Reasoning and Planning with Puzzles](https://mlsanity.com/paper/2510.06475/) (2025)
- [Baba is LLM: Reasoning in a Game with Dynamic Rules](https://mlsanity.com/paper/2506.19095/) (2025)
- [CayleyPy Growth: Efficient growth computations and hundreds of new conjectures on Cayley graphs (Brief version)](https://mlsanity.com/paper/2509.19162/) (2025)
- [From Building Blocks to Planning: Multi-Step Spatial Reasoning in LLMs with Reinforcement Learning](https://mlsanity.com/paper/2512.24532/) (2025)
- [PHANTOM RECALL: When Familiar Puzzles Fool Smart Models](https://mlsanity.com/paper/2510.11812/) (2025)

## Abstract (arXiv)

The application of Generative Pre-trained Transformer (GPT-2) to learn text-archived game notation provides a model environment for exploring sparse reward gameplay. The transformer architecture proves amenable to training on solved text archives describing mazes, Rubik's Cube, and Sudoku solvers. The method benefits from fine-tuning the transformer architecture to visualize plausible strategies derived outside any guidance from human heuristics or domain expertise. The large search space (>1019) for the games provides a puzzle environment in which the solution has few intermediate rewards and a final move that solves the challenge.