princeton-nlp
/

warm-startgrpothink__Llama-3.1-8B-Instruct

Model card Files Files and versions

warm-start__grpo__think__Llama-3.1-8B-Instruct / longcot_config.json

princeton-nlp's picture

Uploading the models

d6d5150 verified about 1 month ago

history blame contribute delete

124 Bytes

	{
	"longcot": true,
	"longcot_delimiter": "</think>",
	"end_delimiter": null,
	"start_think_marker": "<think>"
	}