unseenDem/

convo_001

	convo_001




	
	Dataset Description

This dataset contains conversational data prepared for fine-tuning language models with proper user/assistant turn alternation.

Training conversations: 10657 Validation conversations: 968 Test conversations: 964 Average conversation length: 11.5 turns Format: Each conversation follows the ChatML format with user/assistant roles

Each example contains a conversations field with alternating… See the full description on the dataset page: https://huggingface.co/datasets/unseenDem/convo_001.