Search papers, labs, and topics across Lattice.
10
0
11
5
Current video generation models fail to maintain embodiment consistency and functional interaction in human-to-robot manipulation, revealing significant gaps in their transfer capabilities.
Safety in education-facing LLMs is inversely related to their teaching effectiveness, challenging assumptions about model specialization.
Current vision-language models struggle with process understanding in robotic manipulation, but targeted post-training can yield significant improvements.
Sycophancy fine-tuning can induce severe misalignment in language models, but Alignment Gating offers a powerful solution to reverse this trend while preserving model performance.
Large-scale generative models struggle with low-level vision tasks, revealing critical performance gaps that conventional metrics fail to capture.
Current AI agents struggle to reliably rediscover scientific knowledge, with top performers averaging only 21.5 out of a possible score, revealing critical gaps in their research capabilities.
Reward models that adapt to fine-grained, task-specific criteria can significantly improve text-to-image generation by better aligning with user preferences.
Current video generation benchmarks overlook crucial aspects of physical plausibility and temporal coherence, highlighting the need for holistic evaluation metrics like PhyScore.
MLLMs still struggle to recognize themselves in a mirror, revealing surprising gaps in their self-centric understanding despite advances in other areas.
LLMs can now predict other LLMs' performance with 14% higher accuracy, even when only seeing one or two data points, by blending statistical priors with reasoning.
