Metraipolis › archive › dc_4e40235f2391
RRSI: Regularized Recursive Self-Improvement of Agent Harnesses — arXiv:2609.24972
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.24972
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T06:50:44.814Z
- SHA-256 of the archived text
5c8caa339c62168c109e071486167b6e8b427014ef1a90d59e60239cf0386f7c- Stable URL of this page
https://metraipolis.com/documento/dc_4e40235f2391
How to cite
RRSI: Regularized Recursive Self-Improvement of Agent Harnesses — arXiv:2609.24972 — arXiv (en), https://arxiv.org/abs/2609.24972 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4e40235f2391 (Metraipolis archive dc_4e40235f2391, sha256 5c8caa339c62168c109e071486167b6e8b427014ef1a90d59e60239cf0386f7c)
BibTeX
@misc{dc4e40235f2391,
title = {RRSI: Regularized Recursive Self-Improvement of Agent Harnesses — arXiv:2609.24972},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.24972},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4e40235f2391; sha256 5c8caa339c62168c109e071486167b6e8b427014ef1a90d59e60239cf0386f7c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.24972. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
An LLM agent's capability is largely magnified by its harness, namely the prompts, control flow, tooling, memory, and context management surrounding the frozen backbone model. Recent methods increasingly automate this process by iteratively proposing and selecting component-wise edits of an agent harness, practically establishing a form of recursive self-improvement (RSI) at the agent-system level. However, such recursive evolution may overfit by memorizing the training tasks, showing large in-distribution gains that shrink or even vanish on out-of-distribution benchmarks. We introduce Regularized Recursive Self-Improvement of Agent Harnesses (RRSI), which incorporates the principles of regularizations into harness self-improvement by constraining the evolution candidate proposal and selection. The proposer operates with a temporally annealed budget, limiting how many edits a candidate can bundle, and it encourages unexplored trajectories based on evolution history. The selector is equipped with a critic and a pruner: the critic screens benchmark-specific proposals, while the pruner, removes changes that are too small, too expensive, or no longer useful. Together these constraints favor reusable agent mechanisms over benchmark-specific ones or even noises. Across eight benchmarks spanning coding, agentic workspace and engineering design tasks, RRSI gains up to 6.0 points on the split it evolves against and up to 4.3 points on the five out-of-distribution benchmarks, while producing a harness that runs on 30% fewer policy tokens than the unregularized evolution. Code is available at this https URL and project page is this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CL). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.24972 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-09T06:48:45.158Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).