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.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.21516 [cs.LG] |
| (or arXiv:2605.21516v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.21516 arXiv-issued DOI via DataCite |
From: Boyuan Wang [view email]
[v1]
Fri, 15 May 2026 12:47:13 UTC (2,188 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。