Dataset · O1-OPEN
SFT data for activating the chain of thought
O1-OPEN/OpenO1-SFT
Dataset for language model SFT fine-tuning to generate coherent reasoning chains in English and Chinese.
What's inside
Contains data for language model fine-tuning using SFT to activate Chain-of-Thought. Collected for training the generation of sequential logical reasoning; languages are English and Chinese. Size ranges from 10 to 100k examples; format not specified in the card.
What you can do with it
- LLM SFT fine-tuning for Chain-of-Thought generation
- LoRA training for reasoning tasks
- Testing the model's ability to build logical reasoning
- Tasks
- Text generation
- Languages
- English, Chinese
- Size
- 10K–100K rows
- Downloads this month
- 836
- Likes
- 399
Models trained on it
- reaperdoesntknow/TameForCasualLMText generation 2.9K
- reaperdoesntknow/Qemma-sftText generation 2.6K
- reaperdoesntknow/Qemma-reduxText generation 2.4K
- mradermacher/Llama-3.1-8B-Reason-Blend-888k-i1-GGUF 1.3K
- bartowski/Llama-3.1-8B-Open-SFT-GGUFText generation 1.2K
- mradermacher/Llama-3.1-8B-OpenO1-SFT-v0.1-i1-GGUF 510
- mradermacher/Llama-3.1-8B-Open-SFT-i1-GGUF 465
- mradermacher/llama-3.2-MEDIT-3B-o1-i1-GGUF 372
How to load it
from datasets import load_dataset
ds = load_dataset("O1-OPEN/OpenO1-SFT")