---
tags:
- sudoku
pretty_name: sudoku-extrem
size_categories:
- 1M<n<10M
---
# Dataset Card for Dataset Name

<!-- Provide a quick summary of the dataset. -->

This dataset is built on top of [`sapientinc/sudoku-extreme`](https://huggingface.co/datasets/sapientinc/sudoku-extreme) by preprocessing every example with [`timvink/sudoku-solver`](https://github.com/timvink/sudoku-solver).

The additional fields are:
- `trajectory`: the step-by-step sudoku board configuration from the question to the solution, acquired by running `puzzle.solve_step()` until complete
- `num_steps`: number of calls to `puzzle.solve_step()` required to obtain solution
- `strategies_used`: set of strategies used by the solver