LATIDIA · Investigación

Curar el contexto siempre cargado para los agentes de LLM: un modelo de surtido capacitado con comentarios censurados

arXiv:2610.11007v1 Tipo de anuncio: nuevo Resumen: Al comienzo de cada sesión, los agentes de LLM cargan un archivo de contexto fijo, como $\texttt{AGENTS.md}$. Cada token cargado en el archivo se cobra de nuevo en cada ronda posterior de

La noticia

arXiv:2610.11007v1 Announce Type: new Abstract: At the start of every session, LLM agents load a fixed context file, such as $\texttt{AGENTS.md}$. Each loaded token in the file is charged again in every later round of the session, and these files can degrade performance as they grow in size. However, in practice, human or automated curators usually grow these files by appending. We formulate context curation as a capacitated assortment problem. Instructions consume tokens under a finite attention capacity; adding an instruction never raises the compliance of the others, while retained instructions incur a per-session setup cost. We prove an upper bound on the optimal file size, regardless of the number