AI RESEARCH
LLM-AutoDP: Automatic Data Processing via LLM Agents for Model Fine-tuning
arXiv CS.LG
•
ArXi:2601.20375v2 Announce Type: replace Large Language Models (LLMs) can be fine-tuned on domain-specific data to enhance their performance in specialized fields. However, such data often contains numerous low-quality samples, necessitating effective data processing (DP). In practice, DP strategies are typically developed through iterative manual analysis and trial-and-error adjustment. These processes inevitably incur high labor costs and may lead to privacy issues in high-privacy domains like healthcare due to direct human access to sensitive data.