Sign in to view source links and access this dataset
Description
Reasoning Mix 100k is a high-quality, balanced dataset of 99,999 samples extracted from three reasoning datasets. It is formatted for models to utilize a /think block for step-by-step reasoning, focusing on mathematics, logic, and general problem-solving. The dataset was created by PinkPixel and last updated on 2026-05-11.
Use Cases
Train language models for mathematical problem-solving based on the described mathematics tasks.
Fine-tune models for logical reasoning based on the described logic tasks.
Develop and benchmark step-by-step reasoning capabilities using the /think block format.
Create synthetic data pipelines for generating reasoning traces based on the curated collection of tasks.
Strengths
Contains 99,999 samples, providing a substantial volume for training.
Curated from three high-performing reasoning datasets, suggesting quality and diversity.
Specifically formatted for step-by-step reasoning with a /think block, indicating a clear design purpose.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is known, but the specific source datasets and their individual contributions are not detailed.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
PinkPixel on Hugging Face.
Collection Method
Curated collection extracted from three reasoning datasets.
Freshness
Last updated 2026-05-11 02:53:31; freshness should be verified.
License is unknown; terms of use must be verified.