{
  "id": 6778,
  "url": "https://arxiv.org/abs/2603.23990v1",
  "title": "From Untamed Black Box to Interpretable Pedagogical Orchestration: The Ensemble of Specialized LLMs Architecture for Adaptive Tutoring",
  "summary": "Monolithic Large Language Models (LLMs) used in educational dialogue often behave as \"black boxes,\" where pedagogical decisions are implicit and difficult to audit, frequently violating instructional constraints by providing answers too early. We introduce the Ensemble of Specialized LLMS (ES-LLMS) architecture that separates decision-making from wording. Pedagogical actions are selected by a deterministic rules-based orchestrator coordinating specialized agents covering tutoring, assessment, fe",
  "authors": "Nizam Kadir",
  "category": "research",
  "topics": "agents-autonomy,transparency",
  "orgs": null,
  "regions": null,
  "published_at": "2026-03-25T06:38:19.000Z",
  "fetched_at": "2026-07-14T16:32:45.892Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/6778",
  "original_url": "https://arxiv.org/abs/2603.23990v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}