Metraipolis › archive › dc_15e08daea635
Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM Agents — arXiv:2609.02889
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.02889
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T01:37:15.853Z
- SHA-256 of the archived text
62039632c2b4e4712bf6c04e492860c629cc924a93144b0bfa62cf1d81064690- Stable URL of this page
https://metraipolis.com/documento/dc_15e08daea635
How to cite
Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM Agents — arXiv:2609.02889 — arXiv (en), https://arxiv.org/abs/2609.02889 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_15e08daea635 (Metraipolis archive dc_15e08daea635, sha256 62039632c2b4e4712bf6c04e492860c629cc924a93144b0bfa62cf1d81064690)
BibTeX
@misc{dc15e08daea635,
title = {Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM Agents — arXiv:2609.02889},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.02889},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_15e08daea635; sha256 62039632c2b4e4712bf6c04e492860c629cc924a93144b0bfa62cf1d81064690},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.02889. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A growing body of work improves frozen large language models (LLMs) as agents by evolving their harness: the textual scaffolding around the model, including persona, strategy, format rules, and control heuristics. Existing reflective prompt-evolution methods usually optimize this harness as one flat string. We instead ask where the optimization value actually resides. We introduce HARNESSEVO, which decomposes the harness into four separately evolvable slots: role, task-strategy, tool/format-rules, and reflection/control. Using the same reflective optimizer under an iso-budget setting, we pair this decomposition with leave-one-in and leave-one-out attribution to measure the contribution of each slot. On ALFWorld with a frozen 7B backbone, HARNESSEVO does not significantly improve the overall binary success rate over either the stock harness or flat-string evolution: 0.657 versus 0.642 and 0.642, respectively. However, the slot-level analysis reveals that nearly all useful optimization value is localized in the reflection/control slot, which achieves a leave-one-in gain of +0.119. The other slots are individually null. We further show that uniform budget splitting is harmful: allocating 64 rollouts across four slots leaves only 16 per slot, below the optimizer's effective search floor, causing every slot to freeze at its empty seed. Concentrating the budget on the high-credit control slot recovers the lost gain, reaching 0.761 with half the split budget. The effect is task-contingent. On WebShop, all slots freeze empty and all methods tie, indicating a genuine absence of recurrent, verbalizable control failures rather than budget starvation. Overall, our results suggest that harness value is localized, uniform budget splitting can be actively harmful, and credit assignment should precede structured agent-evolution. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.02889 — 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 2026-10-17T01:11:43.475Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).