LLaVAction: evaluating and training multi-modal large language models for action recognition Paper • 2503.18712 • Published 12 days ago • 3