Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

When Only the Final Text Survives: Implicit Execution Tracing for Multi-Agent Attribution

Yi Nian, Haosen Cao, Shenzhe Zhu, Henry Peng Zou, Qingqing Luan, Yue Zhao

Abstract

When a multi-agent system produces an incorrect or harmful answer, who is accountable if execution logs and agent identifiers are unavailable? Multi-agent language systems increasingly rely on structured interactions such as delegation and iterative refinement, yet the final output often obscures the underlying interaction topology and agent contributions. We introduce IET (Implicit Execution Tracing), a metadata-independent framework that enables token-level attribution directly from generated text and a simple mechanism for interaction topology reconstruction. During generation, agent-specific keyed signals are embedded into the token distribution, transforming the text into a self-describing execution trace detectable only with a secret key. At detection time, a transition-aware scoring method identifies agent handover points and reconstructs the interaction graph. Experiments show that IET recovers agent segments and coordination structure with high accuracy while preserving generation quality, enabling privacy-preserving auditing for multi-agent language systems.

When Only the Final Text Survives: Implicit Execution Tracing for Multi-Agent Attribution

Abstract

When a multi-agent system produces an incorrect or harmful answer, who is accountable if execution logs and agent identifiers are unavailable? Multi-agent language systems increasingly rely on structured interactions such as delegation and iterative refinement, yet the final output often obscures the underlying interaction topology and agent contributions. We introduce IET (Implicit Execution Tracing), a metadata-independent framework that enables token-level attribution directly from generated text and a simple mechanism for interaction topology reconstruction. During generation, agent-specific keyed signals are embedded into the token distribution, transforming the text into a self-describing execution trace detectable only with a secret key. At detection time, a transition-aware scoring method identifies agent handover points and reconstructs the interaction graph. Experiments show that IET recovers agent segments and coordination structure with high accuracy while preserving generation quality, enabling privacy-preserving auditing for multi-agent language systems.
Paper Structure (78 sections, 17 equations, 3 figures, 5 tables)

This paper contains 78 sections, 17 equations, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Background.
  3. Problem Definition.
  4. Our Approach.
  5. Contributions.
  6. Method
  7. Problem Formulation
  8. Attribution via Sequential Scoring.
  9. Robustness to Metadata Loss.
  10. Structural Validation.
  11. Implicit Execution Tracing
  12. Agent Attribution
  13. Local Consistency Scoring.
  14. Competitive Change-Point Detection.
  15. Execution Path and Topology Encoding
  16. ...and 63 more sections

Figures (3)

  • Figure 1: Agent provenance recovering problem in multi-agent systems. Internal execution paths and agent contributions are visible during generation but discarded at delivery due to privacy security, leaving only the final text. The task is to recover latent path boundaries and agent participation from text alone.
  • Figure 2: Overview.
  • Figure 3: Robustness to PII redaction. Attribution performance on original and PII-redacted transcripts across different interaction structures. The similar curves indicate minimal degradation after redaction.