Raja Giryes

A novel gating mechanism allows LLMs to learn new tasks without forgetting old ones, even without access to prior training data.

Joint audio-video diffusion models hallucinate not from diffuse attention noise, but because bidirectional audio-video cross-attention systematically overrides text conditioning in favor of learned canonical priors.

Policy gradient methods may be self-defeating in language model reasoning, as their inherent entropy reduction chokes off exploration and limits downstream performance.