Published November 12, 2026 | Version 1.0.0

Supervisory Capacity Model: dataset for "How Many Agents Can One Supervisor Track? A Mechanistic Capacity Model and Measurement Protocol for LLM-Agent Teams"

  • 1. Clemson University
  • 2. Metacognition Institute

Description

Parameters and outputs of the mechanistic supervisory-capacity model (capacity parameters, service-time variance, simulation curves and summary), the computational probe pilot (per-row results and summary), and the scripts that produce them.

Dataset accompanying the paper How Many Agents Can One Supervisor Track? A Mechanistic Capacity Model and Measurement Protocol for LLM-Agent Teams, Human-AI Coevolution Workshop (HAIC) at NeurIPS 2026 (https://openreview.net/forum?id=czINRnBTpf).

Paper abstract. Agentic deployments can assign many parallel agents to one supervisor without specifying when that supervisor can no longer recover the team's state. Human-factors research already establishes relevant constructs such as situation awareness, span of control, neglect tolerance, and multiple-resource workload, but language-model teams distribute state across text and can fail silently. This position paper translates those constructs into a measurement protocol designed for later calibration to people rather than a new human experiment. Typed event rates, isolated service costs, and a throughput budget define a nominal load-balance index; the midpoint of a separately fitted capacity curve remains an empirical quantity. Synthetic controls recover a known capacity boundary, while a seeded queue shows that deadline pressure can separate the fitted midpoint from nominal load balance. The resulting event ontology, operating envelope, executable stress tests, and preregisterable falsification criteria allow system designers to test staffing or interface interventions within a declared team-monitor channel. They do not turn synthetic behavior into evidence about human or model supervisors.

Known limitations.

  • Simulation and computational-pilot data only; the human study the protocol specifies has not been run.

License: Creative Commons Attribution 4.0 International (https://creativecommons.org/licenses/by/4.0/). Source code files are released under the MIT License.

Files

croissant.json

Files (46.4 kB)

Name Size
md5:9a84a4f2936aea0ef3ffbfbf6573ba2d
1.3 kB Download
md5:4bf657a38651f6d3a50d0eb1449520d8
9.2 kB Preview Download
md5:88ffdf926e8c860fe6996b55583fef02
1.6 kB Download
md5:24fba34892f45524780e9ac820f1e7dc
5.3 kB Preview Download
md5:3375af331db016f5988b0878a34ef14d
28.9 kB Preview Download

Additional details

Related works