Search papers, labs, and topics across Lattice.
6
5
13
11
Uncovering the root cause of modality interference, this work achieves a remarkable 28.5% improvement in full-duplex interaction fluidity without sacrificing efficiency.
Training updates that improve performance in LLMs can actually degrade inference quality鈥攗nless you use the new Monotonic Inference Policy Update framework.
Mainstream models falter in multi-reference image generation, but DyRef's innovative training framework boosts their performance significantly.
Debugging complex code agents just got easier: CodeTracer reconstructs full state transition histories, pinpointing failure origins and enabling recovery of failed runs.
An open-source ecosystem for agentic learning, complete with a trained agent and novel policy optimization, promises to accelerate research by providing a standardized, scalable platform.
MLLMs can be made significantly safer without sacrificing performance by disentangling risks in multimodal inputs and using RLAIF, outperforming even GPT-4V by 16% on safety benchmarks.
