mattwesney/Reasoning_Problem_Solving_Dataset

Reasoning and Problem-Solving Dataset (RPSD)

Synthetic English question–answer pairs for reasoning and problem solving across science, mathematics, engineering, computing, philosophy, and practical scenarios.

Contents

Rows: 151,261
Format: English text; JSON records with question, answer, id, and metadata (difficulty and topic).

Topics: Deductive, inductive, analogical, abductive, and causal reasoning; puzzles, algorithms, planning, and cross-domain problems.

Creation and limitations

The Genesis II pipeline generated the pairs using Qwen, Phi, and Llama models, with structured prompting and format checks.

Some difficulty values incorrectly default to 1, and difficulty labels may occasionally appear in answer text. Synthetic answers can contain errors or biases; review them before use.

Use

from datasets import load_dataset

dataset = load_dataset("mattwesney/Reasoning_Problem_Solving_Dataset")

Request access and authenticate if the dataset is gated.

License: MIT. Version: 1.0.0.

Citation

Matthew Wesney (2025), Reasoning and Problem-Solving Dataset (RPSD). Hugging Face dataset.