Understanding the Effects of Domain Finetuning on LLMs

Eshaan Tanwar, Deepak Nathani, William Yang Wang +1

Large Language Models (LLMs) fine-tuned for specific domains exhibit strong performance; however, the underlying mechanisms by which this fine-tuning reshapes their parametric spac…