Beyond Consensus: Trace-Level Synthesis in Mixture of Agents

• URL: http://arxiv.org/abs/2605.29116v1
• Date: Wed, 27 May 2026 21:24:35 GMT
• Title: Beyond Consensus: Trace-Level Synthesis in Mixture of Agents
• Abstract summary: We show that an aggregator that reads complete reasoning traces recovers correct solutions even when agents unanimously agree.<n>A single model with perturbation-induced trace variation outperforms heterogeneous model pools across structured reasoning, PhD-level science, competition mathematics, and competitive programming.
• Score: 1.0195618602298682
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: When multiple LLM agents solve the same problem, standard practice compresses each agent's reasoning into a majority vote or layered synthesis, treating agreement as the finish line. We show this is unnecessarily lossy: an LLM aggregator that reads complete reasoning traces recovers correct solutions even when agents unanimously agree, with beneficial corrections consistently outweighing harmful ones -- the \emph{aggregation paradox}. Majority voting has a ceiling that perturbation diversity does not raise (error correlations are identical); the aggregator's gain comes from trace-level complementarity, assembling correct intermediate steps from minority chains that voting discards. These findings motivate Self-Consistent Mixture of Agents which generates trace diversity through semantic-preserving input perturbations, safeguards the majority via anchored refinement with provable non-degradation guarantees, and always synthesizes -- never gates on consensus. A single model with perturbation-induced trace variation outperforms heterogeneous model pools across structured reasoning, PhD-level science, competition mathematics, and competitive programming. The unit of aggregation should be the reasoning trace, not the answer.

Related papers

• TwistedMerge: Certified Higher-Order Diagnostics and Abstention for Model Merging [0.48086260459837454]
  We formulate merging as a finite descent problem in which checkpoints are local objects, alignment maps are transitions, and cycle products are residuals.<n>We prove a constant-edge no-go result, frozen-complex three-way and pre-family error-control theorems, and a refinement test for comparison-complex sensitivity.<n>A trained low-rank-adapter audit shows that naive factor averaging depends on the chosen GLr representative, whereas global factor synchronization and dense-delta SVD are stable.
  arXiv  Detail & Related papers  (2026-07-23T03:24:50Z)
• LC-ERD: Mining Latent Logic for Self-Evolving Reasoning via Consistency-Regulated Reward Decomposition [55.572260012037084]
  We introduce LC-ERD (Logic-Consistent Endogenous Reward Decomposition), a framework framing self-alignment as latent structure mining.<n>We derive a Variational Logic Potential by aggregating consensus from the model's Latent Logic Expertise.<n>Experiments show LC-ERD delivers a robust self-evolution path, uncovering trade-offs between logic consistency and accuracy.
  arXiv  Detail & Related papers  (2026-05-19T07:27:50Z)
• \mathsf{VISTA}: Decentralized Machine Learning in Adversary Dominated Environments [21.69919643934826]
  Decentralized machine learning often relies on outsourcing computations, such as evaluations, to untrusted worker nodes.<n>We study this adversary-dominated setting through an incentive-oriented framework in which reports are accepted and rewarded only when they are mutually consistent up to a threshold.<n>We propose mathsfVISTA, an adaptive algorithm that tunes the acceptance threshold using the optimization history.
  arXiv  Detail & Related papers  (2026-05-08T15:07:15Z)
• The Consensus Trap: Rescuing Multi-Agent LLMs from Adversarial Majorities via Token-Level Collaboration [32.17285446393425]
  Response-level aggregation collapses when corrupted agents form a local majority.<n>We propose the Token-Level Round-Robin (RR) Collaboration, where agents sequentially interleave generation within a shared auto-regressive context.<n>We prove that RR maintains robust accuracy well beyond this critical threshold.
  arXiv  Detail & Related papers  (2026-04-18T20:31:37Z)
• Heterogeneous Consensus-Progressive Reasoning for Efficient Multi-Agent Debate [58.675713546748305]
  Multi-Agent Debate (MAD) is a collaborative framework in which multiple agents iteratively refine solutions through the generation of reasoning and alternating critique cycles.<n>This work introduces Heterogeneous Consensus-Progressive Reasoning for Efficient Multi-Agent Debate (HCP-MAD)<n>HCP-MAD employs a three-stage progressive reasoning mechanism to develop adaptive solutions across varying task complexities.
  arXiv  Detail & Related papers  (2026-04-03T06:58:59Z)
• Test-Time Scaling with Diffusion Language Models via Reward-Guided Stitching [66.39914384073145]
  We propose a self-consistency framework that turns cheap diffusion-sampled reasoning into a reusable pool of step-level candidates.<n>We find that step-level recombination is most beneficial on harder problems.<n>Our training-free framework improves average accuracy by up to 2 across six math and coding tasks.
  arXiv  Detail & Related papers  (2026-02-26T11:08:39Z)
• Multi-Agent Debate: A Unified Agentic Framework for Tabular Anomaly Detection [7.807411603618987]
  Multi-Agent Debating framework (MAD) treats disagreement as a first-class signal.<n> MAD produces a normalized anomaly score, confidence, and structured evidence, augmented by a large language model (LLM)-based critic.<n>We show how conformal calibration can wrap the debated score to control false positives under exchangeability.
  arXiv  Detail & Related papers  (2026-02-15T17:44:32Z)
• A Single Revision Step Improves Token-Efficient LLM Reasoning [3.344806691289323]
  We introduce Packet-Conditioned Revision (PACER), a training-free, inference-only framework for large language models.<n>PACER enables reasoning traces to revise their conclusions through a structured coordination step.<n>On challenging competitive math benchmarks, PACER matches or exceeds the accuracy of 256-sample majority voting.
  arXiv  Detail & Related papers  (2026-02-02T21:28:42Z)
• CARE What Fails: Contrastive Anchored-REflection for Verifiable Multimodal [84.71254539482369]
  Group-relative reinforcement learning with verifiable rewards (RLVR) often wastes the most informative data it already has the failures.<n>We present CARE, a failure-centric post-training framework for multimodal reasoning that turns errors into supervision.<n> CARE improves accuracy and training smoothness while explicitly increasing the share of learning signal that comes from failures.
  arXiv  Detail & Related papers  (2025-12-22T16:34:21Z)
• Eigen-1: Adaptive Multi-Agent Refinement with Monitor-Based RAG for Scientific Reasoning [53.45095336430027]
  We develop a unified framework that combines implicit retrieval and structured collaboration.<n>On Humanity's Last Exam (HLE) Bio/Chem Gold, our framework achieves 48.3% accuracy.<n>Results on SuperGPQA and TRQA confirm robustness across domains.
  arXiv  Detail & Related papers  (2025-09-25T14:05:55Z)
• When Disagreements Elicit Robustness: Investigating Self-Repair Capabilities under LLM Multi-Agent Disagreements [56.29265568399648]
  We argue that disagreements prevent premature consensus and expand the explored solution space.<n>Disagreements on task-critical steps can derail collaboration depending on the topology of solution paths.
  arXiv  Detail & Related papers  (2025-02-21T02:24:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.