Skip to content

Dataset · faunix

40k reasoning chains from Qwen3.8-27B

faunix/Qwen3.8-27B-Distillation-40K

Distilled step-by-step reasoning from Qwen3.8-27B across four domains for training thinking models.

What's inside

Contains 40,000 English reasoning chain examples generated locally by the Qwen3.8-27B model in FP8 precision with medium reasoning effort. Prompts for generation were sourced from 12 open datasets covering 4 subject areas. Data is formatted in tabular and text views.

What you can do with it

  • SFT fine-tuning of more compact language models on reasoning skills
  • Distilling logical inference from Qwen3.8-27B into custom models
  • Analyzing step-by-step reasoning structure at a medium level of detail
Tasks
Text generation
Languages
English
Size
10K–100K rows
Downloads this month
1.1K
Likes
43

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("faunix/Qwen3.8-27B-Distillation-40K")