Recent projects highlight how open-source and open-weight models can be refined to achieve frontier-level performance and enhanced safety. AI security company Alice announced results from a project where it collaborated with AI21 Labs and Inflection AI to improve Nvidia Nemotron models. Working with AI21, Alice protected enterprise agents against indirect prompt injection by post-training Nemotron 3.5 Super VL. Simul
Forbes ·