Search for a command to run...
Chain-of-Frames: Advancing Video Understanding in Multimodal LLMs via Frame-Aware Reasoning