The paper introduces a self‑evolving harness framework where a frozen language‑model agent first solves tasks and then edits its own harness based on run records. Using a 49‑line seed harness, the evolved harness improves average scores on in‑distribution benchmarks by 4.48 points and on out‑of‑distribution benchmarks by 12.64 points, surpassing Codex on the former and matching it on the latter. Continued evolution on a specific out‑of‑distribution benchmark further raises performance, and the study analyzes emergent mechanisms such as output truncation and history compaction.
By Qiankai Xu
Automatic Heuristic Design (AHD) has emerged as a transformative approach for solving combinatorial optimization problems. While recent Large Language Model (LLM)-based methods have shown promise, they predominantly rely on fixed evolutionary operators and struggle to effectively accumulate and reuse historical search experience.
arXiv:2610.15150v1 Announce Type: new
Abstract: Recent work on self-evolving LLM agents treats the agent harness as an optimization target that can be iteratively improved through execution feedback....
By Weichen Yu, Lesly Miculicich, Bhavana Dalvi Mishra, Jiefeng Chen, Shuo Li, Matt Fredrikson, Tomas Pfister, Long T. Le
HGM-Dual is a diagnostic‑guided framework that improves harness optimization for planning agents by representing agent evolution as lineage‑aware context and using error‑conditioned dual‑path child generation. It records ancestor changes, behavior outcomes, and ineffective modifications, enabling explicit edit‑outcome attribution. The method achieves up to 14.45% relative improvement in case accuracy while using 2×–4× fewer evaluation budgets compared to the strongest baseline.
By Sudipta Paul, Pengrui Quan, Aounon Kumar, Nicholas Tzou, Vivek Kulkarni, Srinivas Chappidi
The paper introduces EMO-STA, a two‑stage evolutionary multi‑task optimization framework for large‑language‑model (LLM) guided program discovery. First, it evolves a shared archive of executable programs across a family of related tasks; then it adapts selected shared candidates to each target task using strategies such as warm‑starting, best‑average, or best‑per‑task adaptation. Experiments on eight diverse task families show that EMO‑STA outperforms single‑task evolution under matched budgets, with balanced shared and adaptation budgets yielding the best results and shared evolution helping to reduce overfitting in low‑evidence scenarios.
By Halil Alperen Gozeten, Xuechen Zhang, Emrullah Ildiz, Ege Onur Taga, Tara Javidi, Samet Oymak
arXiv:2606. 29082v1 Announce Type: cross Abstract: Would experience designing faster GPU kernels also help close in on a long-standing open mathematical conjecture?
By Young-Jun Lee, Seungone Kim, Minki Kang, Alistair Cheong Liang Chuen, Zerui Chen, Seungho Han, Taehee Jung, Dongyeop Kang