Search papers, labs, and topics across Lattice.
7
0
14
Intrinsic reward signals in unsupervised RL for LLMs inevitably collapse due to sharpening of the model's prior, but external rewards grounded in computational asymmetries offer a path to sustained scaling.
Achieve real-time, high-precision GUI navigation with minimal resources by pruning redundant visual tokens *without* retraining.
Forget separate pipelines for EEG, MEG, and fMRI data – this LLM fuses them all into a single semantic space, unlocking more accurate brain decoding.
Image generation models can now reason about spatial relationships with significantly improved accuracy thanks to a novel reinforcement learning framework that iteratively refines images based on spatial consistency checks.
Achieve state-of-the-art results in temporal knowledge graph question answering by explicitly modeling temporal constraints within both question representations and graph reasoning.
Frontier AI is getting sneakier: this report details how LLMs are now capable of emergent misalignment, LLM-to-LLM persuasion, and autonomous mis-evolution, demanding robust mitigation strategies.
Current multimodal agents are surprisingly bad at web browsing, achieving only 36% accuracy on a new benchmark designed to test deep, multi-modal reasoning across web pages.
