Harnesses for Inference-Time Alignment over Execution Trajectories

• URL: http://arxiv.org/abs/2605.21516v1
• Date: Fri, 15 May 2026 12:47:13 GMT
• Title: Harnesses for Inference-Time Alignment over Execution Trajectories
• Abstract summary: We study harness design through the lens of inference-time trajectory alignment.<n>This decomposition allows us to quantify how workflow, retry budgets, and guidance-induced action reweighting shape the performance limits of harness design.<n>Inspired by the theory, we show that effective harnesses can be partial: specifying only the initial steps and leaving the remaining execution to agent can achieve higher pass rate.
• Score: 13.182534464050695
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Harness engineering has emerged as an important inference-time technique for large language model (LLM) agents, aiming to improve long-term performance through task decomposition and guided execution. However, more elaborate harnesses are not uniformly better: increasing decomposition or guidance can sometimes improve execution, but can also reduce final task success. We study harness design through the lens of inference-time trajectory alignment. This perspective separates harness into two mechanisms: task decomposition, which structures a task into sub-goals, and guided execution, which reshapes local action distributions during execution. This decomposition allows us to quantify how workflow granularity, retry budgets, and guidance-induced action reweighting shape the performance limits of harness design. It further reveals concrete failure modes, including over-decomposition, over-pruning, and hallucinated execution. We validate these predictions through controlled synthetic experiments and real terminal agent benchmarks. Inspired by the theory, we further show that effective harnesses can be partial: specifying only the initial steps and leaving the remaining execution to agent can achieve higher pass rate than fully structured workflows.

Related papers

• Ecdysis: Efficient and Effective Training of Runtime Harnesses for LLM Agents [25.41577484327689]
  Self-evolving runtime harnesses can substantially improve the capabilities of large language model (LLM) agents.<n>Existing harness evolution methods rely on iterative search, repeatedly evaluating and revising candidate harnesses.<n>We propose Ecdysis, an efficient and effective framework that distinguishes model-specific accommodation from systematic harness deficiencies.
  arXiv  Detail & Related papers  (2026-09-10T15:09:08Z)
• Rethink Before You Execute: Adaptive Execution for World Action Models [29.76405877028887]
  World Action Models jointly predict future actions and the evolution of the environment.<n>We argue that this fixed execution horizon is poorly matched to execution dynamics.<n>We propose TempoWAM, a lightweight plug-and-play execution scheme for WAMs.
  arXiv  Detail & Related papers  (2026-08-10T11:59:42Z)
• EASy: Towards Efficient LLM-Based Agentic System [54.10819421625103]
  We propose EASy, a trainable agentic framework that jointly optimize task performance and computational efficiency through reinforcement learning.<n> EASy equips an LLM-based orchestrator with explicit knowledge of the capability and cost profiles of heterogeneous executors, enabling context-sensitive coordination beyond performance-only routing.<n>It further introduces a milestone-plan-act workflow that decomposes complex tasks into manageable milestones, constructs dependency-aware execution graphs, assigns suitable executors, and parallelizes independent steps while adapting subsequent decisions to intermediate outcomes.
  arXiv  Detail & Related papers  (2026-08-05T08:50:36Z)
• Recursive Harness Self-Improvement [28.432258880153125]
  Under model--harness co-evolution, harnesses are data-generating components whose execution traces can shape future foundation models.<n>We introduce Recursive Harness Self-Improvement (RHI), which represents the harness as a prompt-level specification of the agent loop.<n>RHI iteratively refines it using pairwise feedback over its own revision history.
  arXiv  Detail & Related papers  (2026-07-17T00:21:19Z)
• CausalFlow: Causal Attribution and Counterfactual Repair for LLM Agent Failures [4.061481215870679]
  We introduce CausalFlow, an interventional framework that converts failed agent traces into minimal counterfactual repairs and reusable supervision.<n>CaulFlow supports two complementary uses: targeted test-time repair that recovers from failures with minimal behavioral drift, and training-time supervision suitable for offline preference optimization or reward modeling.
  arXiv  Detail & Related papers  (2026-05-25T01:47:01Z)
• TIER: Trajectory-Invariant Execution Rewards for Multi-Step Tool Composition [62.56752617853322]
  Outcome-based rewards provide only sparse feedback, while trajectory-supervised rewards depend on annotated reference solutions.<n>We propose TIER: Trajectory-Invariant Execution Rewards, a reward framework that derives supervision directly from function schemas and runtime execution.
  arXiv  Detail & Related papers  (2026-05-16T03:47:26Z)
• From Reaction to Anticipation: Proactive Failure Recovery through Agentic Task Graph for Robotic Manipulation [45.37258918150899]
  We introduce AgentChord, an agentic system that models a manipulation task as a directed task graph.<n>Before execution, this graph is enriched with anticipatory recovery branches that specify context-aware corrective behaviors.<n> Empirical studies on diverse long-horizon bimanual manipulation tasks demonstrate that AgentChord substantially improves success rates and execution efficiency.
  arXiv  Detail & Related papers  (2026-05-12T11:00:45Z)
• Robust and Efficient Tool Orchestration via Layered Execution Structures with Reflective Correction [55.13278005189741]
  We model tool orchestration as learning a layered execution structure that captures high-level tool dependencies.<n>We introduce a schema-aware reflective correction mechanism that detects and repairs errors locally.<n>This design confines errors to individual tool calls and avoids re-planning entire execution trajectories.
  arXiv  Detail & Related papers  (2026-02-21T22:20:01Z)
• daVinci-Agency: Unlocking Long-Horizon Agency Data-Efficiently [35.39097522391409]
  Large Language Models (LLMs) excel at short-term tasks, scaling them to long-horizon agentic synthesis remains challenging.<n>We propose daVinci-Agency, which systematically mines structured supervision from chain-of-PRs.<n>DaVinci-Agency's PR-grounded structure inherently preserves the causal dependencies and iterative refinements essential for teaching persistent goal-cycle behavior.
  arXiv  Detail & Related papers  (2026-02-02T13:23:39Z)
• Provable Benefit of Curriculum in Transformer Tree-Reasoning Post-Training [76.12556589212666]
  We show that curriculum post-training avoids the exponential complexity bottleneck.<n>Under outcome-only reward signals, reinforcement learning finetuning achieves high accuracy with sample complexity.<n>We establish guarantees for test-time scaling, where curriculum-aware querying reduces both reward oracle calls and sampling cost from exponential to order.
  arXiv  Detail & Related papers  (2025-11-10T18:29:54Z)
• Efficient Diffusion as Low Light Enhancer [63.789138528062225]
  Reflectance-Aware Trajectory Refinement (RATR) is a simple yet effective module to refine the teacher trajectory using the reflectance component of images. textbfReflectance-aware textbfDiffusion with textbfDistilled textbfTrajectory (textbfReDDiT) is an efficient and flexible distillation framework tailored for Low-Light Image Enhancement (LLIE)
  arXiv  Detail & Related papers  (2024-10-16T08:07:18Z)
• Watch Every Step! LLM Agent Learning via Iterative Step-Level Process Refinement [50.481380478458945]
  Iterative step-level Process Refinement (IPR) framework provides detailed step-by-step guidance to enhance agent training. Our experiments on three complex agent tasks demonstrate that our framework outperforms a variety of strong baselines.
  arXiv  Detail & Related papers  (2024-06-17T03:29:13Z)

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.