Closing the Consistency Gap: Self-Evolving Agents That Learn to Stay on Course

• URL: http://arxiv.org/abs/2609.08832v1
• Date: Tue, 08 Sep 2026 14:53:43 GMT
• Title: Closing the Consistency Gap: Self-Evolving Agents That Learn to Stay on Course
• Abstract summary: Large language model (LLM)-powered agents can be accurate on average yet unreliable in production.<n>We present a self-evolving framework that reduces this gap by identifying unstable, low-consistency steps in agent trajectories.<n>On AppWorld with ReAct/GPT-4.1, our framework raises the fraction of tasks that succeed in all five runs by +16 points on same-task evaluation and +13 points on similar-task generalization.
• Score: 5.9756241643216255
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Large language model (LLM)-powered agents can be accurate on average yet unreliable in production, a discrepancy that has been observed but remains largely unaddressed. When given the same task five times, a ReAct agent on the AppWorld benchmark using GPT-4.1 succeeds in all five runs only 53% of the time, even though its per-run pass rate averages 77%. We call this 24-point shortfall the consistency gap, and we argue that addressing it is a precondition for trustworthy AI agent deployment. We present a self-evolving agent framework that reduces this gap by identifying unstable, low-consistency steps in agent trajectories and converting them into episodic memory the agent can draw on in future runs. At its core is a Consistency Analyzer that pinpoints where and why a trajectory is likely to flip across executions, and a Guideline Generator that converts the diagnosis into targeted guidelines, committed to memory and injected into future agent executions on similar tasks. On AppWorld with ReAct/GPT-4.1, our framework raises the fraction of tasks that succeed in all five runs by +16 points on same-task evaluation and +13 points on similar-task generalization.

Related papers

• AgentAudit: An Open, Extensible Framework for Full-Lifecycle Trust Evaluation of AI Agents [2.8000808592516293]
  AgentAudit evaluates the entire execution trace across ten capability, grounding, security and behavioural dimensions.<n>It reads only the recorded execution trace and does not interfere with how the agent runs.<n>We evaluate five language models (OpenAI GPT-5, Claude Sonnet 5, Sarvam 105B, Llama 3.3 70B and Gemini 2.5 Flash) across nine capability and adversarial tasks.
  arXiv  Detail & Related papers  (2026-09-09T08:29:16Z)
• PAST-Bench: Benchmarking the Foundations of Recursive Self-Improvement in Personal Agents [50.9061759859778]
  PAST-Bench is a benchmark designed to isolate whether retained experience actually improves agents over time.<n>We report both later-task gains and whether those gains follow the intended save, retrieve, and update pathway.<n>We develop Hermes+, which extends Hermes with five targeted interventions across stages of the agent loop.
  arXiv  Detail & Related papers  (2026-08-04T17:58:05Z)
• AgentAbstain: Do LLM Agents Know When Not to Act? [19.223843408018997]
  Existing evaluations mostly focus on task success rather than whether agents know when to abstain.<n>We present the first systematic evaluation framework for agentic abstention: the calibrated ability of tool-using LLM agents to recognize when not to act.<n>AgentAbstain is a paired-task benchmark built on an agent-native taxonomy of 8 abstention scenarios across pre-execution reasoning and runtime discovery.
  arXiv  Detail & Related papers  (2026-07-11T00:57:36Z)
• ACCORD: Action-Conditioned Contextual Grounding for Language Agents [69.73525608707764]
  ACCORD (Action-Conditioned Contextual Grounding) is an agent framework for adaptive grounding.<n>We show that current agents often fail to do so. They act from assumed rather than observed specifics, overlook information they could have gathered, and fail to incorporate evidence that has already been returned.
  arXiv  Detail & Related papers  (2026-06-15T09:05:55Z)
• The Meta-Agent Challenge: Are Current Agents Capable of Autonomous Agent Development? [80.24951682268332]
  We introduce the Meta-Agent Challenge (MAC), an evaluation framework designed to test the capacity of frontier models for autonomous agent development.<n>To ensure evaluation integrity, this framework is secured by multi-layer defenses against reward hacking.<n>We demonstrate that meta-agents rarely match human-engineered baseline policies, and the few that do are dominated by proprietary frontier models.
  arXiv  Detail & Related papers  (2026-06-03T04:58:17Z)
• EvoCode-Bench: Evaluating Coding Agents in Multi-Turn Iterative Interactions [9.297494684604763]
  EvoCode-Bench is a benchmark of 26 stateful coding tasks and 227 evaluated rounds.<n>Each task preserves the agent's workspace for 5-15 rounds, states requirements through observable behavior.<n>We evaluate 13 coding agents with two metrics: MT@4, a four-attempt fail-stop multi-round score, and SR, a single-round score from a reference-completed prior state.
  arXiv  Detail & Related papers  (2026-05-22T18:17:28Z)
• Shepherd: A Runtime Substrate Empowering Meta-Agents with a Formalized Execution Trace [37.66254228361724]
  We introduce Shepherd, a functional programming model that formalizes meta-agent operations on target agents as functions.<n>Shepherd records every agent-environment interaction as a typed event in a Git-like execution trace, enabling any past state to be forked and replayed.<n>The system forks the agent process and its $5$ faster than Docker, achieving $>95%$ prompt-cache reuse on replay.
  arXiv  Detail & Related papers  (2026-05-11T17:50:51Z)
• 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)
• AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agents [50.481033105867205]
  We introduce AgentProcessBench, the first benchmark dedicated to evaluating step-level effectiveness in realistic, tool-augmented trajectories.<n>The benchmark comprises 1,000 diverse trajectories and 8,509 human-labeled step annotations with 89.1% inter-annotator agreement.<n>It features a ternary labeling scheme to capture exploration and an error propagation rule to reduce labeling ambiguity.
  arXiv  Detail & Related papers  (2026-03-15T16:13:58Z)
• Towards a Science of Scaling Agent Systems [79.64446272302287]
  We formalize a definition for agent evaluation and characterize scaling laws as the interplay between agent quantity, coordination structure, modelic, and task properties.<n>We derive a predictive model using coordination metrics, that cross-validated R2=0, enabling prediction on unseen task domains.<n>We identify three effects: (1) a tool-coordination trade-off: under fixed computational budgets, tool-heavy tasks suffer disproportionately from multi-agent overhead, and (2) a capability saturation: coordination yields diminishing or negative returns once single-agent baselines exceed 45%.
  arXiv  Detail & Related papers  (2025-12-09T06:52:21Z)
• Alita-G: Self-Evolving Generative Agent for Agent Generation [54.49365835457433]
  We present ALITA-G, a framework that transforms a general-purpose agent into a domain expert.<n>In this framework, a generalist agent executes a curated suite of target-domain tasks.<n>It attains strong gains while reducing computation costs.
  arXiv  Detail & Related papers  (2025-10-27T17:59:14Z)

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.