Model Security

Researchers are searching for potential backdoors in Jane Street's large language models (LLMs), citing concerns about model safety and reliability. The investigation is ongoing, with no concrete findings reported yet.