LGAIMay 15

Harnesses for Inference-Time Alignment over Execution Trajectories

arXiv:2605.2151694.8
Predicted impact top 4% in LG · last 90 daysOriginality Incremental advance
AI Analysis

Provides a theoretical framework and practical insights for designing more effective inference-time harnesses for LLM agents, addressing a known bottleneck in agent performance.

The paper studies inference-time harness design for LLM agents, separating it into task decomposition and guided execution. It identifies failure modes like over-decomposition and shows that partial harnesses (specifying only initial steps) can outperform fully structured workflows.

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.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes