Harnesses for Inference-Time Alignment over Execution Trajectories
This paper studies harness engineering for LLM agents, an inference-time technique that uses task decomposition and guided execution. It reveals that more elaborate harnesses are not always better, identifying failure modes like over-decomposition, over-pruning, and hallucinated execution. Experiments show that partial harnesses, which specify only initial steps, can achieve higher pass rates than fully structured workflows.
[2605.21516] Harnesses for Inference-Time Alignment over Execution Trajectories
[Submitted on 15 May 2026]
Title:Harnesses for Inference-Time Alignment over Execution Trajectories
View a PDF of the paper titled Harnesses for Inference-Time Alignment over Execution Trajectories, by Boyuan Wang and 4 other authors
View PDF HTML (experimental)
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
Submission history
From: Boyuan Wang [view email] [v1] Fri, 15 May 2026 12:47:13 UTC (2,188 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled Harnesses for Inference-Time Alignment over Execution Trajectories, by Boyuan Wang and 4 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.LG
new | recent | 2026-05
Change to browse by:
cs cs.AI
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
Loading...
Data provided by:
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)