README.md

metadata

language:
  - en
license: apache-2.0
task_categories:
  - conversational
  - text-generation
tags:
  - reasoning
  - distillation
  - sft
  - deepseek
  - kimi
  - qwen
size_categories:
  - 10K<n<100K

distill-sft-26k

A merged SFT dataset of 25,973 high-quality reasoning conversations distilled from three frontier models.

Source Datasets

Source	Rows	Reasoning	Domain
Spakie/DeepSeek-V4-Pro-distilled	17,670	No	General
trjxter/Kimi-K2.6-Reasoning-3300x-WandB	3,303	Yes (all, ``)	General / Science / Math
WithinUsAI/Qwen3.7_Max_Thinking_dataset_5K	5,000	Yes (all)	Math / Science

Format

Standard chat format with optional reasoning:

{
  "messages": [
    {"role": "user", "content": "..."},
    {"role": "assistant", "content": "...", "reasoning_content": "..."}
  ],
  "source": "deepseek_v4_pro | kimi_k2_6 | qwen3_7_max",
  "domain": "general | science | math",
  "id": "..."
}

• reasoning_content present on 8,303 rows (32%)
• Shuffled with seed=42

Loading

from datasets import load_dataset
ds = load_dataset("GenueAI/distill-sft-26k", split="train")

License

Apache 2.0. See source dataset licenses for details.