Paper Detail

Poor Man's Agentic Modeling: Simulating Large LLM-Agent Societies on a Laptop

Igor Itkin

huggingface Score 13.0

Published 2026-07-19 · First seen 2026-08-13

General AI

Abstract

Simulating societies of many large language model (LLM) agents is expensive, yet the questions asked of such simulations are usually macroscopic: phase behaviour, stylised facts, and scaling with the number of agents N, not the cognition of any single agent. We turn a statistical-physics observation into a method: replace each LLM agent by a low-parameter model fitted from a few hundred to a few thousand cheap queries, then run the society at any N on a laptop. Whether this works is decided before the simulation runs, chiefly by what each agent perceives. We introduce an [interaction order x memory] taxonomy that maps perception and memory to an effective theory and a predicted N-trend of the surrogate error. We validate it on a faithful reimplementation of the LLM macroeconomy EconAgent and seven further named LLM simulations, with agent decisions cloned from genuine LLM elicitations (primarily DeepSeek) for a few dollars; the predicted error trends hold cell by cell, and the two refuted predictions, both on a strongly saturating response and traced to its curvature, are themselves matched quantitatively by the theory with no free parameters.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@misc{itkin2026poor,
  title = {Poor Man's Agentic Modeling: Simulating Large LLM-Agent Societies on a Laptop},
  author = {Igor Itkin},
  year = {2026},
  abstract = {Simulating societies of many large language model (LLM) agents is expensive, yet the questions asked of such simulations are usually macroscopic: phase behaviour, stylised facts, and scaling with the number of agents N, not the cognition of any single agent. We turn a statistical-physics observation into a method: replace each LLM agent by a low-parameter model fitted from a few hundred to a few thousand cheap queries, then run the society at any N on a laptop. Whether this works is decided befo},
  url = {https://huggingface.co/papers/2608.11215},
  keywords = {large language model agents, low-parameter surrogate models, statistical physics, interaction order, memory taxonomy, effective theory, N-scaling, EconAgent, DeepSeek, surrogate error, code available, huggingface daily},
  eprint = {2608.11215},
  archiveprefix = {arXiv},
}

Metadata

{}