Search papers, labs, and topics across Lattice.
6
0
11
YouZhi-LLM achieves unprecedented concurrency and accuracy in financial LLMs by dramatically reducing KV-cache overhead, setting a new standard for deployment efficiency.
Even the best MLLMs struggle to meet user requirements, achieving only 66% coverage of essential task functions.
Low-probability tokens can be the key to distinguishing AI-generated text from human writing, revealing hidden distributional discrepancies that traditional methods overlook.
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.
Simple adaptive negative sampling can significantly boost the performance of knowledge graph foundation models on zero-shot completion tasks.
By explicitly conditioning on the query, QGS achieves a 0.62% CTR increase in a major commercial search engine, proving that generative models can beat traditional deep learning baselines in search ranking when query context is properly handled.
