Beyond Symmetric Agents: Cognitive Diversity and Multi-Agent Debate in Small Language Models
Source
Leonardo Ferreira, Gardenia Liu, Kaden Zheng
Author
Leonardo Ferreira, Gardenia Liu, Kaden Zheng
Date
Key takeaways · AI-distilled
The study spans 23 small open weightsA model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.Full definition → models from eleven vendor families, five tasks and more than 5,500 debate and control runs, pairing every debate setup with a majority-vote control given the same generation budget.
Debate beat single-AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → inference by 3 to 7 points on tasks with headroom, but at matched budgets it tied or lost to self-consistency sampling while using 1.6x the wall-clock time and 3.4x the tokens.
Persona prompting lowered accuracy. A dose-response test over each model's persona combinations points to a persona tax, not a diversity tax: redundant personas hurt most, and maximally diverse teams recovered part of the loss.
Mixed-model teams lost to majority votes over their own rosters, with accuracy tracking member capability rather than heterogeneity, and nearly all of debate's benefit came from the first exchange of answers.
The authors flag a measurement hazard: debate transcripts silently overflowed serving context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → windows, and correcting that alone moved their debate-versus-sampling gap from -1.8 points to parity.
Terms in this piece · Glossary
multi-agent — Using several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.
open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
Why it matters
If you are adding debate or persona agents, budget-matched majority voting often matches them at far lower cost. This tests the assumption that agent diversity drives multi-agentUsing several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.Full definition → gains.
articleMulti-Agent Debate Strategies: Survey, Taxonomy, and ChallengesQuim Motger, Marc Oriol, Jordi Marco, Xavier Franch
articleThe Wisdom of Artificial Deliberative CrowdsFederico Barrera-Lemarchand, Mariano Sigman, Joaquin Navajas
articleWhy Better Models Can Create Riskier Systems: Evidence from LLM Agents in Financial MarketsJillian Ross, Eric So, Zoe De Simone, Charles Pozniak, Andrew W. Lo