Agent / long-term memory

MemGuard

MemGuard: Preventing Memory Contamination in Long-Term Memory-Augmented Large Language Models

On the frontierfirst seen May 27, 2026

Current frontier — recent, not yet superseded in the knowledge base

0 papers critique it · 0 beat it on benchmarks

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.