-
Orca: Progressive Learning from Complex Explanation Traces
The Orca approach involves leveraging explanation tuning to generate detailed responses from a large language model. -
Evol-Instruct: A Pipeline for Automatically Evolving Instruction Datasets
The Evol-Instruct pipeline involves automatically evolving instruction datasets using large language models. -
LaMini: A Large-Scale Instruction Dataset
The LaMini approach involves generating a large-scale instruction dataset by leveraging the outputs of a large language model, gpt-3.5-turbo.
