LATIDIA · Investigación

MemCut: Cada secreto deja un rastro en la memoria a largo plazo de los agentes de LLM

arXiv: 2610.21405v1Tipo de anuncio: nuevo Resumen: La memoria a largo plazo se ha convertido en un componente crucial de los agentes del modelo de lenguaje grande (LLM), y la mayoría de las investigaciones se centran en recuperar información relevante con precisión y eficiencia

La noticia

arXiv:2610.21405v1 Announce Type: new Abstract: Long-term memory has become a crucial component of large language model (LLM) agents, with most research focusing on retrieving relevant information accurately and efficiently. However, a dependable memory system must also forget sensitive user data upon request. To evaluate this capability, we introduce MemCut, a systematic benchmark for user-requested forgetting. Its 456 hand-designed secrets in long-horizon conversations have fully annotated evidence trails to test recovery after deletion, including from messages that never state the secrets directly. We propose two recovery metrics. Native recovery measures reader success through the system's retrieval, whereas residual recovery measures the same reader's success when directly given surviving objects traced to