Trace-Level Analysis of Information Contamination in Multi-Agent Systems

• URL: http://arxiv.org/abs/2604.27586v1
• Date: Thu, 30 Apr 2026 08:39:42 GMT
• Title: Trace-Level Analysis of Information Contamination in Multi-Agent Systems
• Abstract summary: Reasoning over heterogeneous artifacts increasingly occurs within structured agent.<n>We study this phenomenon by treating uncertainty as a controlled variable.<n>We characterize three manifestation types: silent semantic corruption, behavioral detours with recovery, and combined structural disruption and control-flow signatures.
• Score: 8.829951903444352
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reasoning over heterogeneous artifacts (PDFs, spreadsheets, slide decks, etc.) increasingly occurs within structured agent workflows that iteratively extract, transform, and reference external information. In these workflows, uncertainty is not merely an input-quality issue: it can redirect decomposition and routing decisions, reshape intermediate state, and produce qualitatively different execution trajectories. We study this phenomenon by treating uncertainty as a controlled variable: we inject structured perturbations into artifact-derived representations, execute fixed workflows under comprehensive logging, and quantify contamination via trace divergence in plans, tool invocations, and intermediate state. Across 614 paired runs on 32 GAIA tasks with three different language models, we find a decoupling: workflows may diverge substantially yet recover correct answers, or remain structurally similar while producing incorrect outputs. We characterize three manifestation types: silent semantic corruption, behavioral detours with recovery, and combined structural disruption and their control-flow signatures (rerouting, extended execution, early termination). We measure operational costs and characterize why commonly used verification guardrails fail to intercept contamination. We contribute (i) a formal taxonomy of contamination manifestations in structured workflows, (ii) a trace-based measurement framework for detecting and localizing contamination across agent interactions, and (iii) empirical evidence with implications for targeted verification, defensive design, and cost control.

Related papers

• Constraint-Aware Synthetic Tabular Data Generation via Inter-Column Constraint Discovery with LLM Agents [3.7680093320209553]
  We study the discovery and enforcement of three complementary inter-column constraint families.<n>Our unified tool-grounded workflow represents all three as machine-executable hypotheses.
  arXiv  Detail & Related papers  (2026-08-15T08:19:54Z)
• ScrambleToolBench: Agents Search Exhaustively Even When Their Own Map Points to the Next Step [74.83535434786145]
  We introduce ScrambleToolBench, an interactive terminal benchmark designed to isolate behavioral reasoning.<n>The benchmark requires agents to uncover hidden tool behaviors entirely through trial-and-error interaction.<n>Our evaluation of state-of-the-art language models reveals that successful initial discovery does not translate into robust adaptation.
  arXiv  Detail & Related papers  (2026-08-03T15:07:46Z)
• Monitoring Agentic Systems Before They're Reliable [0.7422381022274628]
  Agentic systems entering production typically operate as partially integrated assemblies where structural defects, not task-level errors, dominate the failure landscape.<n>We present a monitoring and triage methodology that decomposes agentic system evaluation into three dimensions (quality, suitability, efficiency) at three monitoring scopes.<n>Findings are routed through severity classification adapted from FMEA, concentrating human attention on the subset that warrants investigation.
  arXiv  Detail & Related papers  (2026-06-01T17:01:53Z)
• Meta-Agent: From Task Descriptions to Verified Multi-Agent Systems [36.485119937780524]
  We present Meta-Agent, a two-phase framework that automatically constructs and executes specialized multi-agent systems.<n>We introduce a three-level error attribution mechanism that distinguishes local, upstream, and structural failures.<n>Experiments against strong multi-agent baselines and ablation studies demonstrate consistent improvements in task success rate, error recovery, and workflow stability.
  arXiv  Detail & Related papers  (2026-05-24T19:38:54Z)
• LiveFMBench: Unveiling the Power and Limits of Agentic Workflows in Specification Generation [75.05397479715576]
  Large language models (LLMs) and agents have shown promising progress, but their true capabilities and failure modes remain unclear.<n>We present the first systematic and contamination-aware study of LLM- and agent-based formal specification generation for C programs.
  arXiv  Detail & Related papers  (2026-05-02T11:31:33Z)
• From Perception to Autonomous Computational Modeling: A Multi-Agent Approach [0.0]
  We present a solver-agnostic framework in which coordinated large language model (LLM) agents autonomously execute the complete computational mechanics workflow.<n>Agents are formalised as conditioned operators on a shared context space with quality gates that introduce conditional iteration between pipeline layers.
  arXiv  Detail & Related papers  (2026-04-08T07:56:34Z)
• Beyond Fluency: Toward Reliable Trajectories in Agentic IR [0.0]
  This position paper synthesizes failure modes observed in industrial agentic systems.<n>We argue that safe deployment requires moving beyond endpoint accuracy toward trajectory integrity and causal attribution.
  arXiv  Detail & Related papers  (2026-04-05T21:20:57Z)
• Dynamic analysis enhances issue resolution [53.50448142467294]
  DAIRA (Dynamic Analysis-enhanced Issue Resolution Agent) is an automated repair framework that embeds dynamic analysis into the agent's reasoning cycle.<n> Driven by a Test Tracing-Driven methodology, DAIRA utilizes lightweight monitors to extract critical runtime data.<n>When powered by Gemini 3 Flash Preview, DAIRA establishes a new state-of-the-art (SOTA) performance, achieving a 79.4% resolution rate on the SWE-bench Verified dataset.
  arXiv  Detail & Related papers  (2026-03-23T14:48:54Z)
• Dynamics Within Latent Chain-of-Thought: An Empirical Study of Causal Structure [58.89643769707751]
  We study latent chain-of-thought as a manipulable causal process in representation space.<n>We find that latent-step budgets behave less like homogeneous extra depth and more like staged functionality with non-local routing.<n>These results motivate mode-conditional and stability-aware analyses as more reliable tools for interpreting and improving latent reasoning systems.
  arXiv  Detail & Related papers  (2026-02-09T15:25:12Z)
• Noise & pattern: identity-anchored Tikhonov regularization for robust structural anomaly detection [58.535473924035365]
  Anomaly detection plays a pivotal role in automated industrial inspection, aiming to identify subtle or rare defects in otherwise uniform visual patterns.<n>We tackle structural anomaly detection using a self-supervised autoencoder that learns to repair corrupted inputs.<n>We introduce a corruption model that injects artificial disruptions into training images to mimic structural defects.
  arXiv  Detail & Related papers  (2025-11-10T15:48:50Z)
• Traceability and Accountability in Role-Specialized Multi-Agent LLM Pipelines [9.340611077939828]
  We study a traceable and accountable pipeline, meaning a system with clear roles, structured handoffs, and saved records.<n>We evaluate eight configurations of three state-of-the-art LLMs on three benchmarks and analyze where errors start, how they spread, and how they can be fixed.
  arXiv  Detail & Related papers  (2025-10-08T23:26:44Z)
• Tracking the risk of a deployed model and detecting harmful distribution shifts [105.27463615756733]
  In practice, it may make sense to ignore benign shifts, under which the performance of a deployed model does not degrade substantially. We argue that a sensible method for firing off a warning has to both (a) detect harmful shifts while ignoring benign ones, and (b) allow continuous monitoring of model performance without increasing the false alarm rate.
  arXiv  Detail & Related papers  (2021-10-12T17:21:41Z)
• Cross-Supervised Joint-Event-Extraction with Heterogeneous Information Networks [61.950353376870154]
  Joint-event-extraction is a sequence-to-sequence labeling task with a tag set composed of tags of triggers and entities. We propose a Cross-Supervised Mechanism (CSM) to alternately supervise the extraction of triggers or entities. Our approach outperforms the state-of-the-art methods in both entity and trigger extraction.
  arXiv  Detail & Related papers  (2020-10-13T11:51:17Z)

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.