Game-RL: Synthesizing Multimodal Verifiable Game Data to Boost VLMs' General Reasoning
iclr26-game-rl.github.io
We propose Game-RL, constructing diverse game tasks for RL training to boost VLMs
立即前往
返回首页
