GPT-OSS-120B-Distilled-Reasoning-math: Chain-of-Thought Math Problem Dataset

GPT-oss-120B-Distilled-Reasoning-math is a dataset of mathematical problems with generated reasoning processes and answers. The data was created by author Jackrong using the gpt-oss-120b model and was last updated on August 17, 2025. The dataset is formatted in JSON Lines and includes fields for the question, category, reasoning steps, and final answer.

Use Cases

• Fine-tuning language models for mathematical problem-solving based on the provided reasoning processes.
• Benchmarking model performance on multi-step reasoning tasks based on the structured question-answer pairs.
• Studying the characteristics of distilled chain-of-thought reasoning from a large teacher model.
• Developing educational tools that require step-by-step solutions for math problems.

Strengths

• Includes generated complete reasoning processes, providing step-by-step logic for each problem.
• Contains multiple structured fields such as Category, Input, CoT_Native_Reasoning, Reasoning, and Answer.
• Data was generated using a specific large language model (gpt-oss-120b), indicating a consistent source.

Limitations

• Row count, file size, and column-level documentation are unknown, limiting suitability assessment.
• License information is unavailable, which may restrict usage.
• The dataset's quality and potential biases depend on the generating model and are not independently verified.

Provenance

Source

Author Jackrong on Hugging Face, generated using the gpt-oss-120b model.

Collection Method

Model-generated distillation of reasoning processes for mathematical problems.

Time Range

null

Freshness

Last updated 2025-08-17 07:56:30.

Geography

null