Latte
PulseAugur coverage of Latte — every cluster mentioning Latte across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
New backdoor vulnerability discovered in LLM multi-agent systems
Researchers have identified a new vulnerability in LLM-based multi-agent systems where collaboration can inadvertently trigger backdoor behavior. This occurs when collective evidence from multiple agents reaches a hidde…
-
New caching methods accelerate diffusion model inference, reducing latency up to 6.7x
Researchers have developed new methods to accelerate diffusion model inference by intelligently caching and reusing intermediate features. OnlineCache learns dynamic caching policies and error correction to adapt resour…
-
New framework enhances DNN testing with latent space mutation
Researchers have developed Latte, a new black-box testing framework for deep neural networks designed to improve the identification of model weaknesses. Latte operates by mutating inputs within the network's latent spac…
-
New method uses tables to detect AI images with limited data
Researchers have developed a novel method for detecting AI-generated images, particularly in low-data scenarios where traditional detectors struggle. This approach transforms images into a tabular format, using a frozen…
-
LATTE framework forecasts user preferences for personalized LLM generation
Researchers have introduced LATTE, a novel framework for personalizing large language models (LLMs) by forecasting user preference trajectories. Unlike existing methods that aggregate user history into static profiles, …
-
LATTE framework boosts LLM team efficiency with adaptive task graphs
Researchers have developed a new framework called LATTE to improve the efficiency of large language model (LLM) teams. LATTE addresses inefficiencies in current LLM coordination methods by enabling teams to collaborativ…