Assessing the Macro and Micro Effects of Random Seeds on Fine-Tuning Large Language Models
This addresses the overlooked issue of random seed effects in fine-tuning LLMs, which is crucial for researchers and practitioners to ensure reliable model evaluation and reproducibility, though it is incremental in highlighting an existing problem.
The study systematically evaluated the impact of random seeds on fine-tuning large language models using GLUE and SuperGLUE benchmarks, revealing significant variance in performance metrics like accuracy and F1, as well as in prediction consistency across runs.
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 systematically evaluate the effects of random seeds on LLMs using the GLUE and SuperGLUE benchmarks. We analyze the macro-level impact through traditional metrics like accuracy and F1, calculating their mean and variance to quantify performance fluctuations. To capture the micro-level effects, we introduce a novel metric, consistency, measuring the stability of individual predictions across runs. Our experiments reveal significant variance at both macro and micro levels, underscoring the need for careful consideration of random seeds in fine-tuning and evaluation.