{
  "id": 339,
  "url": "https://arxiv.org/abs/2607.01523v1",
  "title": "Multi-Head Recurrent Memory Agents",
  "summary": "Recurrent memory agents extend LLMs to arbitrarily long contexts by iteratively consolidating input into a fixed-size memory window. Despite their scalability, these agents exhibit a well-documented reliability problem: end-to-end performance degrades systematically as context length grows. We diagnose this failure by decomposing performance into two factors--memory capture and memory retention--and quantitatively confirm that retention is the dominant bottleneck. Retention collapses because exi",
  "authors": "Jiatong Li, Samuel Yeh, Sharon Li",
  "category": "research",
  "topics": "healthcare,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-07-01T22:38:54.000Z",
  "fetched_at": "2026-07-14T14:14:28.436Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/339",
  "original_url": "https://arxiv.org/abs/2607.01523v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}