Coverage, Not Credit: Failure-Credit Routing of Zeroth-Order Perturbation Budgets Does Not Improve On-Pool Sample Efficiency for LLM Agents

• URL: http://arxiv.org/abs/2608.28011v1
• Date: Fri, 28 Aug 2026 07:27:02 GMT
• Title: Coverage, Not Credit: Failure-Credit Routing of Zeroth-Order Perturbation Budgets Does Not Improve On-Pool Sample Efficiency for LLM Agents
• Abstract summary: Trajectory-level credit assignment can localize which module causes failures using only verifiable signals.<n>We ask whether failure credit should route a fixed zeroth-order/evolution-strategies (ZO/ES) perturbation budget.<n>Across a synthetic environment and frozen Qwen2.5-1.5B/3B and SmolLM2-1.7B agents, we find no statistically detectable improvement over uniform allocation in any on-pool comparison.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Trajectory-level credit assignment can localize which module of a tool-using LLM agent causes failures using only verifiable signals. We ask whether such failure credit should route a fixed zeroth-order/evolution-strategies (ZO/ES) perturbation budget. Across a synthetic environment and frozen Qwen2.5-1.5B/3B and SmolLM2-1.7B agents, three task families, six allocation schemes, a credit-noise sweep, paired seeds, and exact sign-flip tests, we find no statistically detectable improvement over uniform allocation in any on-pool comparison (no gain of at least 2 percentage points). The joint soft-plus-sigma scheme is equivalent to uniform within a +/- 0.02 AUC margin on 1.5B and 3B; concentrating the full budget on the credit argmax is marginally equivalent on 1.5B, where that module is the verified bottleneck, and significantly worse on 3B. Inverse-propensity debiasing does not rescue routing, and misrouting costs up to -0.074 AUC in-house and -0.118 end-to-end on the BFCL-derived family. Across six fixed-step schedules, loss is linear in bottleneck starvation rate (R^2 = 0.94, descriptive), and a preregistered credit-free coverage floor removes detected harm. Matched-budget burst and step-compensating catch-up schedules are consistent with harm arising from insufficient cumulative parameter movement rather than update frequency. Our primary estimand is optimization efficiency on a fixed task pool. On unseen BFCL functions, the study's one exception is that soft routing exceeds uniform on held-out endpoints (+0.047, p = 0.031, n = 6). A plausible but untested reading is that routing-favored caller improvements transfer while uniform's on-pool gains reflect a synthesizer behavior specific to our harness. We report this exception explicitly and document three failure modes that can silently invalidate ZO/ES experiments on frozen LLMs.

Related papers

• Accurate Ensembles, Fragile Narratives: Multi-Scale Stacking and a Fidelity Audit of LLM-Generated Explanations for Credit Risk [0.0]
  Credit scoring increasingly relies on models whose decision logic cannot be read off their parameters.<n>A common proposal closes that gap with a language model: compute feature attributions, hand them to an LLM, and let it write the rationale.<n>We build such a system end to end and test whether the second half of the promise holds.
  arXiv  Detail & Related papers  (2026-08-08T13:22:14Z)
• ZEBRA: Zero-shot Budgeted Resource Allocation for LLM Orchestration [8.226365534099399]
  We propose ZEBRA, a framework that reduces multi-phase budget allocation to a continuous nonlinear knapsack problem.<n>On a $150$-task APPS coding benchmark, both ZEBRA variants outperform LLM-direct on every aggregate metric.<n>We show that lightweight algorithmic guidance at inference time can improve the economic behavior of autonomous multi-agent systems.
  arXiv  Detail & Related papers  (2026-05-19T20:50:05Z)
• Unsolvability Ceiling in Multi-LLM Routing: An Empirical Study of Evaluation Artifacts [0.07812854697536452]
  We present a large-scale study of multi-tier LLM routing with 206,000 query-model pairs across six benchmarks.<n>We show that a substantial portion of reported unsolvability stems from evaluation artifacts.
  arXiv  Detail & Related papers  (2026-05-08T07:49:24Z)
• LLM Readiness Harness: Evaluation, Observability, and CI Gates for LLM/RAG Applications [51.56484100374058]
  We present a readiness harness for LLM and RAG applications that turns evaluation into a deployment decision workflow.<n>The system combines automated benchmarks, OpenTelemetry observability, and CI quality gates under a minimal API contract.<n>We evaluate the harness on ticket-routing and BEIR tasks with full Azure matrix coverage.
  arXiv  Detail & Related papers  (2026-03-28T18:03:32Z)
• MOSAIC: Multi-Objective Slice-Aware Iterative Curation for Alignment [1.356919241968803]
  We study how to allocate a fixed supervised fine-tuning budget when three objectives must be balanced at once.<n>We propose MOSAIC, a multi-objective framework for closed-loop data mixture search built on a unified L1-L3 evaluation interface.
  arXiv  Detail & Related papers  (2026-03-19T09:00:47Z)
• 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)
• Trajectory Guard -- A Lightweight, Sequence-Aware Model for Real-Time Anomaly Detection in Agentic AI [0.0]
  Trajectory Guard is a Siamese Recurrent Autoencoder with a hybrid loss function that jointly learns task-trajectory alignment via contrastive learning and sequential validity via reconstruction.<n>At 32 ms latency, our approach runs 17$-27times$ faster than LLM Judge baselines, enabling real-time safety verification in production deployments.
  arXiv  Detail & Related papers  (2026-01-02T00:27:11Z)
• 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)
• On GRPO Collapse in Search-R1: The Lazy Likelihood-Displacement Death Spiral [59.14787085809595]
  We identify Lazy Likelihood Displacement (LLD) as the core mechanism driving this failure.<n>LLD emerges early and triggers a self-reinforcing LLD Death Spiral, where declining likelihood leads to low-confidence responses.<n>We propose a lightweight likelihood-preserving regularization LLDS for GRPO that activates only when a trajectory's likelihood decreases.
  arXiv  Detail & Related papers  (2025-12-03T19:41:15Z)
• Doubly Robust Distributionally Robust Off-Policy Evaluation and Learning [59.02006924867438]
  Off-policy evaluation and learning (OPE/L) use offline observational data to make better decisions. Recent work proposed distributionally robust OPE/L (DROPE/L) to remedy this, but the proposal relies on inverse-propensity weighting. We propose the first DR algorithms for DROPE/L with KL-divergence uncertainty sets.
  arXiv  Detail & Related papers  (2022-02-19T20:00:44Z)
• Label Distributionally Robust Losses for Multi-class Classification: Consistency, Robustness and Adaptivity [55.29408396918968]
  We study a family of loss functions named label-distributionally robust (LDR) losses for multi-class classification. Our contributions include both consistency and robustness by establishing top-$k$ consistency of LDR losses for multi-class classification. We propose a new adaptive LDR loss that automatically adapts the individualized temperature parameter to the noise degree of class label of each instance.
  arXiv  Detail & Related papers  (2021-12-30T00:27:30Z)

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.