Brainquiver/

instruct-narrate-stories-finetune

Instruction-response pairs for instruction finetuning of a small storytelling language model pretrained on Microsoft's TinyStories. The pretrained base already generates stories; this dataset teaches it to follow instructions and respond in a turn-taking format, building on the story distribution learned during pretraining.

	Configs




	

Config Rows Note Schema

deepseek-v4-flash-284b-a13b-high 343,324… See the full description on the dataset page: https://huggingface.co/datasets/Brainquiver/instruct-narrate-stories-finetune.