Math SFT Solutions No CoT: Supervised Fine-Tuning Dataset for Mathematical Problem-Solving

A cleaned mathematics supervised fine-tuning dataset built by kaushik-harsh-99 and last updated on 2026-05-17. It contains instruction-solution pairs, mathematical proofs, derivations, and olympiad-style solutions for theorem reasoning and stepwise explanations. The dataset is designed specifically for mathematical supervised fine-tuning and removes explicit chain-of-thought tags.

Use Cases

• Fine-tuning language models for mathematical problem-solving based on instruction-solution pairs.
• Training models to generate stepwise mathematical explanations based on detailed final solutions.
• Developing systems for theorem reasoning and proof generation based on the included mathematical proofs.
• Creating models capable of solving olympiad-style problems based on the dataset's olympiad-style solutions.

Strengths

• Dataset is specifically built for mathematical supervised fine-tuning (SFT).
• Content includes high-quality mathematical proofs, derivations, and olympiad-style solutions.
• The release removes explicit chain-of-thought tags while preserving solution quality.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count, columns, and file formats are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

huggingface, author kaushik-harsh-99

Collection Method

Likely curated or assembled from mathematical problem sources for supervised fine-tuning.

Freshness

Last updated 2026-05-17 05:17:40; freshness should be verified.