Metraipolis › archive › dc_889e42d0df96
Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training — arXiv:2611.04277
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.04277
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-04T21:37:50.976Z
- SHA-256 of the archived text
097217c2bcadbc9312858bc746472083346bf1fbb2e2d13eb1f7bae7eedf0f92- Stable URL of this page
https://metraipolis.com/documento/dc_889e42d0df96
How to cite
Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training — arXiv:2611.04277 — arXiv (en), https://arxiv.org/abs/2611.04277 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_889e42d0df96 (Metraipolis archive dc_889e42d0df96, sha256 097217c2bcadbc9312858bc746472083346bf1fbb2e2d13eb1f7bae7eedf0f92)
BibTeX
@misc{dc889e42d0df96,
title = {Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training — arXiv:2611.04277},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.04277},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_889e42d0df96; sha256 097217c2bcadbc9312858bc746472083346bf1fbb2e2d13eb1f7bae7eedf0f92},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04277. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Harness evolution from a common base exhibits Locality, Complementarity, and Nonstationarity: variants repair different failures, and their utility changes as the policy learns. Rather than seeking a universally superior harness, we introduce Harness-S to selectively rescue failures of a base execution system through sibling harnesses and transfer the resulting capabilities into a single policy that remains effective under the original harness. Directional Harness Evolution develops specialized sibling harnesses along distinct repair directions, forming a reusable Sibling Harness Pool that remains fixed during policy training. Complementary Harness Rescue Training invokes this pool only when all initial attempts under the base harness fail. The same policy generates fresh rollouts under selected auxiliary harnesses, and these trajectories replace the failed group for reinforcement learning. An Adaptive Harness Router selects the Top-k variants using Harness Behavior Profiles and voting across multiple LLMs. Policy-aware profile re-anchoring refreshes this evidence as the policy learns. Across AppWorld, SWE-Gym, and DBBench, the trained policy substantially outperforms policy-only training when returned to the original base harness. Behavioral analyses and held-out cases support the transfer of harness-elicited behaviors into the policy itself. Reusing the router for Rank-1 inference provides additional gains at the same executor rollout count. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2611.04277 — 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 2026-11-04T21:35:49.530Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).