Search papers, labs, and topics across Lattice.
6
0
10
3
FADE achieves a remarkable 90.4% performance retention in unconstrained open-ended tasks, far surpassing the 48.1% retention of GPT-5.6, revolutionizing counterfactual video understanding.
IntHQ's innovative architecture not only mitigates common pitfalls in multi-task learning but also delivers a measurable 1.60% lift in user engagement for travel recommendations at scale.
LongHorizon-Harness boosts LLM performance on long-horizon tasks by up to 28.9% by revolutionizing task state management and error correction.
Multi-dimensional Evaluation-Verification Reward transforms multi-reference image editing by providing a structured approach to evaluate and enhance visual consistency, yielding superior results over existing models.
LLMs can handle basic route planning, but fall apart when user preferences enter the mix, as shown by a new benchmark based on real-world queries.
Key contribution not extracted.
