Graph Feedback Controls Consensus and Clique Formation in Open-Weight Language-Model Populations
Quick Answer
The study investigates convention formation in open-weight language models (1.1B-32B parameters) using a naming-game protocol, revealing that while retained partner-label evidence is necessary, it alone does not ensure consensus.
Quick Take
The study investigates convention formation in open-weight language models (1.1B-32B parameters) using a naming-game protocol, revealing that while retained partner-label evidence is necessary, it alone does not ensure consensus. Notably, Qwen2.5-32B achieved stable behavioral consensus across all settings, contrasting with threshold-similarity methods that failed to reach consensus in 189 trials.
Key Points
- Open-weight models show varied consensus behavior based on interaction graph dynamics.
- Qwen2.5-32B consistently achieved stable behavioral consensus in all tested settings.
- Threshold-similarity routing often leads to fragmentation rather than consensus.
- Retained history generally shifts fragmented dynamics toward consensus in homogeneous populations.
- Early-window graph-energy features serve as diagnostics for model behavior.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Multi-agent language-model systems increasingly route local interactions, yet the runtime interaction graph is often treated as an implementation detail. We study convention formation in open-weight LM populations spanning 1.1B-32B parameters with a naming-game protocol. Restricted first-token scores over tokenizer-safe labels let us measure prompt-conditioned score-state distributions, construct state-similarity graphs, and separate sampled-label agreement from latent state-space consensus. Across controlled interventions, in the main open-weight repair grids, retained partner-label evidence is necessary but not sufficient: homophilous threshold-similarity routing deletes cross-basin exposure and amplifies fragmentation, while bridge-seeking routing often repairs fragmentation when memory is available. In a three-seed mixed four-model grid, threshold-similarity produces no final behavioral or state consensus in 189 setting-seed runs, whereas state-component and label-disagreement bridges recover final behavioral consensus in 14/18 retained-memory runs. Across homogeneous model populations, retained history generally shifts fragmented dynamics toward consensus; the clearest case is Qwen2.5-32B, which reaches stable behavioral and final state consensus in all 18 retained-history well-mixed settings, while threshold-similarity reaches neither form of consensus in 189 settings. Robustness over state thresholds, population size, and vocabulary size preserves the qualitative ordering, and early-window graph-energy features provide useful within-grid diagnostics.
| Subjects: | Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2607.12077 [cs.AI] |
| (or arXiv:2607.12077v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.12077 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Samer Saab Jr [view email]
[v1]
Mon, 13 Jul 2026 18:55:03 UTC (736 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →The Emerging Paradigm of Geospatial Foundation Models: From Pre-Training to Agentic Reasoning
The paper introduces Geospatial Foundation Models (GeoFMs), AI/ML models pre-trained on extensive geospatial datasets, enabling domain experts to fine-tune them for specific tasks. This paradigm shift democratizes access to advanced AI/ML while ensuring security, and proposes a framework for cost-effective adaptation strategies. The vision of Agentic Geospatial Reasoning is also presented, where Large Language Models orchestrate GeoFMs to automate complex analytical workflows.