Assessing the Macro and Micro Effects of Random Seeds on Fine-Tuning Large Language Models

Nghia Bui, Guergana Savova, Lijing Wang

The impact of random seeds in fine-tuning large language models (LLMs) has been largely overlooked despite its potential influence on model performance.In this study, we systematic…