# Emergent Convergence in Multi-Agent LLM Annotation

arXiv: https://arxiv.org/abs/2512.00047 · HTML: https://mlsanity.com/paper/2512.00047/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use.
- Builds on 8 of its 24 references that are in the map.
- First posted on arXiv 17 Nov 2025.

## Placed in

- LLM agents and tool use / [General LLM agent frameworks and benchmarks](https://mlsanity.com/agents/general-llm-agent-frameworks-and-benchmarks/)

## Builds on (most cited first, 8 of 8)

- [On the Opportunities and Risks of Foundation Models](https://mlsanity.com/paper/2108.07258/) (2021)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [SummaC: Re-Visiting NLI-based Models for Inconsistency Detection in Summarization](https://mlsanity.com/paper/2111.09525/) (2021)
- [Estimating the intrinsic dimension of datasets by a minimal neighborhood information](https://mlsanity.com/paper/1803.06992/) (2018)
- [Improving Language Model Negotiation with Self-Play and In-Context Learning from AI Feedback](https://mlsanity.com/paper/2305.10142/) (2023)
- [Social Influence and the Collective Dynamics of Opinion Formation](https://mlsanity.com/paper/1311.3475/) (2013)
- [Geometric Signatures of Compositionality Across a Language Model's Lifetime](https://mlsanity.com/paper/2410.01444/) (2024)
- [A Computational Method for Measuring "Open Codes" in Qualitative Analysis](https://mlsanity.com/paper/2411.12142/) (2024)

## Built on by (most cited first, 2 of 2)

- [BOUNDARY_SYNC: Measuring Communication-Induced Representational Coupling in Multi-Agent LLM Systems](https://mlsanity.com/paper/2607.01600/) (2026)
- [When Outputs Disperse, Does Epistemic Revision Follow? A Black-Box Diagnostic for Machine Collectives](https://mlsanity.com/paper/2608.03722/) (2026)

## Abstract (arXiv)

Large language models (LLMs) are increasingly deployed in collaborative settings, yet little is known about how they coordinate when treated as black-box agents. We simulate 7500 multi-agent, multi-round discussions in an inductive coding task, generating over 125000 utterances that capture both final annotations and their interactional histories. We introduce process-level metrics: code stability, semantic self-consistency, and lexical confidence alongside sentiment and convergence measures, to track coordination dynamics. To probe deeper alignment signals, we analyze the evolving geometry of output embeddings, showing that intrinsic dimensionality declines over rounds, suggesting semantic compression. The results reveal that LLM groups converge lexically and semantically, develop asymmetric influence patterns, and exhibit negotiation-like behaviors despite the absence of explicit role prompting. This work demonstrates how black-box interaction analysis can surface emergent coordination strategies, offering a scalable complement to internal probe-based interpretability methods.