When Agents Look the Same: Quantifying Distillation-Induced Similarity in Tool-Use Behaviors

• URL: http://arxiv.org/abs/2604.21255v1
• Date: Thu, 23 Apr 2026 03:48:56 GMT
• Title: When Agents Look the Same: Quantifying Distillation-Induced Similarity in Tool-Use Behaviors
• Abstract summary: Existing metrics fail to distinguish mandatory behaviors required for task success from non-mandatory patterns that reflect a model's autonomous preferences.<n>We propose two complementary metrics to isolate non-mandatory behavioral patterns: textbfResponse Pattern Similarity (RPS) for verbal alignment and textbfAction Graph Similarity (AGS) for tool-use habits modeled as directed graphs.
• Score: 66.18091962164219
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Model distillation is a primary driver behind the rapid progress of LLM agents, yet it often leads to behavioral homogenization. Many emerging agents share nearly identical reasoning steps and failure modes, suggesting they may be distilled echoes of a few dominant teachers. Existing metrics, however, fail to distinguish mandatory behaviors required for task success from non-mandatory patterns that reflect a model's autonomous preferences. We propose two complementary metrics to isolate non-mandatory behavioral patterns: \textbf{Response Pattern Similarity (RPS)} for verbal alignment and \textbf{Action Graph Similarity (AGS)} for tool-use habits modeled as directed graphs. Evaluating 18 models from 8 providers on $τ$-Bench and $τ^2$-Bench against Claude Sonnet 4.5 (thinking), we find that within-family model pairs score 5.9 pp higher in AGS than cross-family pairs, and that Kimi-K2 (thinking) reaches 82.6\% $S_{\text{node}}$ and 94.7\% $S_{\text{dep}}$, exceeding Anthropic's own Opus 4.1. A controlled distillation experiment further confirms that AGS distinguishes teacher-specific convergence from general improvement. RPS and AGS capture distinct behavioral dimensions (Pearson $r$ = 0.491), providing complementary diagnostic signals for behavioral convergence in the agent ecosystem. Our code is available at https://github.com/Syuchin/AgentEcho.

Related papers

• Agent trajectories as programs: fingerprinting and programming coding-agent behavior [0.49316866264940024]
  Benchmark scores tell you what an agent got right; they do not tell you how it got there.<n>In this work, we introduce methods for comparing agents procedurally in different contexts, where the model, tasks, and approaches vary.<n>A probe over these procedural signatures attributes an unseen trajectory to the correct agent at 85.7% accuracy, controlling for leakage across tasks.
  arXiv  Detail & Related papers  (2026-06-15T17:28:41Z)
• When Do LLM Agents Treat Surface Noise Differently from Semantic Noise? A 68-Cell Measurement Study with a Held-Out Trace-Level Validation [9.055086193088083]
  We document an empirical phenomenon in chain-of-thought and ReAct agents driven by ten large language models.<n>Mean-bearing perturbations alter final answers more often than presentation perturbations of comparable severity.
  arXiv  Detail & Related papers  (2026-05-25T15:57:11Z)
• How to Interpret Agent Behavior [56.59836196946289]
  We introduce ACT*ONOMY, a taxonomy for describing and analyzing agent behavior at runtime.<n>By providing a shared vocabulary, ACT*ONOMY helps researchers, agent designers, and end users interpret agent behavior more consistently.
  arXiv  Detail & Related papers  (2026-05-13T14:52:40Z)
• Collective Alignment in LLM Multi-Agent Systems: Disentangling Bias from Cooperation via Statistical Physics [0.0]
  We present a model-agnostic statistical-physics method to disentangle social conformity from intrinsic bias, compute critical exponents, and probe the collective behavior and possible phase transitions of multi-agent systems.<n>We find that collective alignment is dominated by an intrinsic bias rather than by cooperative neighbor coupling, producing field-driven crossovers instead of genuine phase transitions.
  arXiv  Detail & Related papers  (2026-05-11T13:13:44Z)
• Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment [59.536125286960186]
  Multi-agent frameworks assigning specialized roles are increasingly adopted to enable self-reflection and mutual auditing.<n>We find it simultaneously induces a human-like cognitive bias known as Actor-Observer Asymmetry.<n>ReTAS guides agents to synthesize conflicting viewpoints into an objective consensus.
  arXiv  Detail & Related papers  (2026-04-21T15:05:58Z)
• TRAJEVAL: Decomposing Code Agent Trajectories for Fine-Grained Diagnosis [23.834704102474927]
  Code agents can resolve GitHub issues, yet when they fail, current evaluation provides no visibility into where or why.<n>We introduce TRAJEVAL, a diagnostic framework that decomposes agent trajectories into three interpretable stages.<n>We validate that these diagnostics are predictive, achieving model-level Pass@1 prediction within 0.87-2.1% MAE.
  arXiv  Detail & Related papers  (2026-03-25T05:27:03Z)
• Descent-Guided Policy Gradient for Scalable Cooperative Multi-Agent Learning [14.185814237633958]
  Descent-Guided Policy Gradient (DG-PG) is a framework that constructs noise-free per-agent guidance gradients.<n>We prove that DG-PG reduces gradient variance from $(N)$ to $mathcalO(1)$, preserves the equilibria of the cooperative game, and achieves agent-independent sample complexity.
  arXiv  Detail & Related papers  (2026-02-23T17:45:08Z)
• Information Fidelity in Tool-Using LLM Agents: A Martingale Analysis of the Model Context Protocol [69.11739400975445]
  We introduce the first theoretical framework for analyzing error accumulation in Model Context Protocol (MCP) agents.<n>We show that cumulative distortion exhibits linear growth and high-probability deviations bounded by $O(sqrtT)$.<n>Key findings include: semantic weighting reduces distortion by 80%, and periodic re-grounding approximately every 9 steps suffices for error control.
  arXiv  Detail & Related papers  (2026-02-10T21:08:53Z)
• From Features to Actions: Explainability in Traditional and Agentic AI Systems [8.859406164948718]
  We bridge the gap between static and agentic explainability by comparing attribution-based explanations with trace-based diagnostics.<n>Our results show that trace-based diagnostics for agentic settings consistently localizes behaviour breakdowns.
  arXiv  Detail & Related papers  (2026-02-06T16:34:29Z)
• GenAgent: Scaling Text-to-Image Generation via Agentic Multimodal Reasoning [54.42973725693]
  We introduce GenAgent, unifying visual understanding and generation through an agentic multimodal model.<n>GenAgent significantly boosts base generator(FLUX.1-dev) performance on GenEval++ and WISE.<n>Our framework demonstrates three key properties: 1) cross-tool generalization to generators with varying capabilities, 2) test-time scaling with consistent improvements across interaction rounds, and 3) task-adaptive reasoning that automatically adjusts to different tasks.
  arXiv  Detail & Related papers  (2026-01-26T14:49:04Z)
• From Harm to Help: Turning Reasoning In-Context Demos into Assets for Reasoning LMs [58.02809208460186]
  We revisit this paradox using high-quality traces from DeepSeek-R1 as demonstrations.<n>We find that adding more exemplars consistently degrades accuracy, even when demonstrations are optimal.<n>We introduce Insight-to-solve (I2S), a sequential test-time procedure that turns demonstrations into explicit, reusable insights.
  arXiv  Detail & Related papers  (2025-09-27T08:59:31Z)
• Reliability, Embeddedness, and Agency: A Utility-Driven Mathematical Framework for Agent-Centric AI Adoption [0.0]
  We formalize three axioms for sustained adoption of agent-centric AI systems executing multi-step tasks.<n>We model adoption as a sum of a decaying novelty term and a growing utility term.
  arXiv  Detail & Related papers  (2025-08-18T12:53:38Z)
• Act-With-Think: Chunk Auto-Regressive Modeling for Generative Recommendation [49.45822979879046]
  Generative recommendation (GR) typically encodes behavioral or semantic aspects of item information into discrete tokens.<n>We present Chunk AutoRegressive Modeling (CAR), a new generation paradigm following the decision pattern that users usually think semantic aspects of items.
  arXiv  Detail & Related papers  (2025-06-30T09:13:54Z)

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.