Clinical Safety Benchmark (CSB)

Neurosymbolic Alignment boosts clinical LLM safety by 21%

Researchers have developed a novel training framework called Neurosymbolic Alignment to enhance the safety of clinical language models. This method integrates a 7B parameter clinical LLM with a physiological world model…