Search papers, labs, and topics across Lattice.
8
1
10
19
Native memory in foundation models can significantly enhance efficiency and performance, as shown by the innovative Metis architecture.
Current LLMs may appear accurate, but they often rely on inconsistent memory states that traditional benchmarks fail to reveal.
No single memory architecture is best for all tasks; performance hinges on how well memory structures align with specific workload challenges.
Small initialization can dramatically enhance reasoning performance in large language models, revealing a new lever for improving AI capabilities.
SparseX achieves efficient KV Cache sharing for LLMs, restoring contextual interactions without the overhead of additional models or preprocessing.
LLM agents can achieve state-of-the-art performance in dynamic environments by treating memory as a continuously evolving graph, rather than a static repository.
Active lifting reveals a novel pathway to enhance slow thinking models, bridging cognitive theory and practical AI applications.
Stop blindly transcribing everything into agent memory: MemReader selectively writes memories based on reasoning, slashing noise and boosting performance on temporal reasoning and knowledge updates.
