研究揭示多智能体 LLM 团队中领导力协调控制的收益边界
Leadership as Coordination Control: Behavioral Signatures and the Recovery-Advantage Boundary in Multi-Agent LLM Teams
AI 导读
arXiv 论文（发表于 TMLR）将交易型、变革型、情境型三种领导风格实现为多智能体 LLM 团队的显式协调控制器，用行为签名和逐动作消融检验其何时有价值。
正文 · 原文
Abstract:Team science holds that leadership is contingent: it helps only under specific conditions. We ask the same question of multi-agent LLM teams: under what measurable conditions does process-level coordination control add value, and do those conditions match what team science predicts? This takes measurement, not just accuracy: we use behavioral signatures (majority lock-in, exploration, recovery from an incorrect round-0 consensus) and per-action ablations, clean because each controller is an explicit action set. We operationalize three classical leadership styles (transactional, transformational, situational) as controllers over a shared action vocabulary (explore, revise, accept, synthesize). A matched controller with the same actions but an arbitrary rule recovers no better than majority voting. Across four task regimes and three open-weight model families, no controller dominates by accuracy. Transactional control matches a shared round-0 vote on all 12 combinations to within 1.3pp, and accuracy gains appear only on the single llama-4-scout social combination, where the round-0 majority is unreliable. Against the stronger flat baseline, only situational still gains (+8pp). A recovery-advantage account, tested with four boundary probes, says a controller beats plain interaction only where the round-0 majority is unreliable, the task is recoverable, and undirected interaction does not already repair it. Because these conditions map onto contingency theory (leadership substitutes, path-goal redundancy, the readiness gap), a largely null accuracy result is what the theory predicts, not a failure of the controllers. We advance this readiness-gap boundary as an explanatory hypothesis rather than an established general principle. Its null side is corroborated across every other combination, including under held-out prediction, while its positive side is observed in a single combination.
| Comments: | Published in Transactions on Machine Learning Research (TMLR), 2026. 37 pages. Reviews: this https URL. Code: this https URL. Video: this https URL |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2606.19111 [cs.CL] |
| (or arXiv:2606.19111v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.19111 arXiv-issued DOI via DataCite |
|
| Journal reference: | Transactions on Machine Learning Research, 2026 |
Submission history
From: Haewoon Kwak [view email]
[v1]
Wed, 17 Jun 2026 14:27:18 UTC (120 KB)
[v2]
Sat, 24 Oct 2026 03:46:30 UTC (182 KB)
来源：arXiv cs.AI · arxiv.org