Dataset · kaist-ai
Chain of Thought Collection
kaist-ai/CoT-Collection
1 million to 10 million English examples with chains of thought for generation and classification.
What's inside
The CoT-Collection by KAIST contains English texts with chains of thought; size ranges from 1 million to 10 million examples. The card specifies text-generation and text-classification tasks, a CC BY 4.0 license, and multiple loading configurations.
What you can do with it
- SFT of a language model on chains of thought
- Training a text classifier with CoT
- Testing a model for explanation generation
- Tasks
- Text generation, Text classification
- Languages
- English
- Size
- 1M–10M rows
- Downloads this month
- 3K
- Likes
- 173
Models trained on it
- mradermacher/b1ade-1b-bf16-i1-GGUF 1.3K
- mradermacher/Reyna-CoT-4B-v0.1-i1-GGUF 1K
- mradermacher/Cypher-CoT-1.8B-i1-GGUF 758
- mradermacher/Cypher-CoT-1.8B-GGUF 423
- mradermacher/Cypher-Laser-Mixtral-2x1.8B-v0.1-GGUF 262
- Bayhakl1/MiniGrimmText generation 238
- mradermacher/Reyna-CoT-4B-v0.1-GGUF 201
- mradermacher/b1ade-1b-bf16-GGUF 199
How to load it
from datasets import load_dataset
ds = load_dataset("kaist-ai/CoT-Collection")