LILI YU

Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better

Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Model

CAT: Content-Adaptive Image Tokenization

LMFusion: Adapting Pretrained Language Models for Multimodal Generation

Jointly Training Large Autoregressive Multimodal Models

Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context Length

Eliciting Attributions from LLMs with Minimal Supervision