Search papers, labs, and topics across Lattice.
4
18
7
3
The DeepSeek-V4.1-Flash model, a multimodal Mixture-of-Experts model with 552B backbone parameters and support for contexts of up to one million tokens, is introduced, substantially improving cost efficiency for agentic workloads and pushing the limits of KV cache compression.
MM-Future, a world-action model that generates multiple paired scene-action hypotheses and models bidirectional interaction within each pair, shows consistent improvements over both single-mode and action-only variants, validating the benefit of multi-mode joint world-action modeling.
This work found that certain attention blocks naturally form an Intrinsic Spatial Grounding Map that precisely locates reference subjects, and proposes Dual-phase Intrinsic Attention Leveraging (DIAL), a framework that uses these internal signals for both training and inference.
Ditch the high-fidelity simulator: IRL-VLA uses a lightweight reward world model trained with inverse reinforcement learning to enable efficient and effective closed-loop RL training for autonomous driving.
