Search papers, labs, and topics across Lattice.
4
0
9
Long-horizon agents can now achieve GPT-6 Astra and Claude Opus-level performance on complex tasks by learning from dense, progressive rewards derived from latent world specifications.
This work adds one feature-alignment term to ordinary VLA training: a frozen world model is run over the training frames once and cached, and the student learns to agree with that cache, indicating a broad representational prior rather than a fragile alignment between two particular networks.
MLLMs struggle with multi-hop reasoning in document-style images, achieving only 62.83% accuracy compared to over 90% for human annotators.
