NVDRS

LLMs outperform fine-tuned models on rare suicide circumstances

A new research paper compares the performance of large language models (LLMs) against fine-tuned RoBERTa models for extracting complex circumstances from death investigation narratives. The study introduces a "Complexit…