The Delegation Danger Band: Why Mid-Capability Sub-Agents Over-Trust Inherited Stale State

cs.MA, cs.AI

Submitted: 2026-09-03

Updated: 2026-09-03

Comments: Preprint. Under review at a NeurIPS 2026 workshop. 16 pages, 3 figures, 9 tables

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Sources

• Lost in the Middle: How Language Models Use Long Contexts
• Entity-Based Knowledge Conflicts in Question Answering
• ConflictBank: A Benchmark for Evaluating the Influence of Knowledge Conflicts in LLM
• MuSiQue: Multihop Questions via Single-hop Question Composition
• Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts
• HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering

Related papers

• Highway Congestion Reduction through Reinforcement Learning Based Eulerian Headway Control
• ORACLE-SWE: Quantifying the Contribution of Critical Information Signals on SWE Agents
• You Only Align Once: Propagating Cooperative Behaviors in Multi-Agent Systems through Seed Agents
• Deny Without Disabling: Authorization-Paired Evaluation and Control for Multi-Agent Systems
• MA-SAPO: Multi-Agent Reasoning for Score-Aware Prompt Optimization
• PeroMAS: A Multi-agent System of Perovskite Material Discovery