A unified supervised fine-tuning dataset merged from 6 high-quality sources. It is formatted into a single messages column in standard chat format and was built for training small reasoning-capable language models. The dataset was created by author 'wop' and was last updated on June 3, 2026.
Use Cases
- Supervised fine-tuning of language models based on the unified chat-formatted data.
- Training small models for reasoning tasks based on the described high-quality sources.
- Benchmarking reasoning capabilities in language models based on the merged dataset structure.
Strengths
- Dataset merges 6 high-quality sources into a single format.
- Data is formatted into a standard chat structure with roles like 'system' and 'user'.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- huggingface
- Collection Method
- Merged from 6 high-quality sources.
- Freshness
- Last updated 2026-06-03 10:11:00; freshness should be verified.