GLM-5.1-Reasoning-1M-Cleaned: Unified SFT-Ready Training Data

GLM-5.1-Reasoning-1M-Cleaned is a cleaned and reformatted derivative of the Kassadin88/GLM-5.1-1000000x dataset, prepared by EngMuhammadAtef. It preserves the original four-subset layout (main, PHD-Science, Multilingual-STEM, Math) while converting examples into a unified schema with explicit conversation, input, output, domain, and meta fields.

Use Cases

• Fine-tuning language models for instruction-following based on the unified SFT-ready schema.
• Training models on specialized reasoning tasks based on the PHD-Science and Math subsets.
• Developing multilingual STEM question-answering systems based on the Multilingual-STEM subset.

Strengths

• Derived from a source dataset containing 1,000,000 examples.
• Organized into four distinct subsets: main, PHD-Science, Multilingual-STEM, and Math.
• Data is reformatted into a unified schema with explicit fields for model training.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update timestamp is 2026-04-21 00:35:22.

Provenance

Source

Derivative of Kassadin88/GLM-5.1-1000000x dataset.

Collection Method

Cleaned and reformatted from the original source.

Freshness

Last updated 2026-04-21 00:35:22.