Metraipolis › archive › dc_a8efa4e063c1
Harness Compilation: Which Decisions Should a Small Vision-Language Model Keep? — arXiv:2610.11231
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.11231
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T05:47:32.633Z
- SHA-256 of the archived text
1fa1f907ce155e81e76c5d3fe5e27f454a6ca27dff198540542335914237bb37- Stable URL of this page
https://metraipolis.com/documento/dc_a8efa4e063c1
How to cite
Harness Compilation: Which Decisions Should a Small Vision-Language Model Keep? — arXiv:2610.11231 — arXiv (en), https://arxiv.org/abs/2610.11231 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a8efa4e063c1 (Metraipolis archive dc_a8efa4e063c1, sha256 1fa1f907ce155e81e76c5d3fe5e27f454a6ca27dff198540542335914237bb37)
BibTeX
@misc{dca8efa4e063c1,
title = {Harness Compilation: Which Decisions Should a Small Vision-Language Model Keep? — arXiv:2610.11231},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.11231},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a8efa4e063c1; sha256 1fa1f907ce155e81e76c5d3fe5e27f454a6ca27dff198540542335914237bb37},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11231. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Small vision-language models may be able to read external evidence yet struggle to obtain it. We introduce Harness Compilation (HC), an offline procedure that adapts the division of work between a frozen small VLM and its external harness. A large teacher uses student execution traces to revise reusable content and control, while a separate validation set selects the deployed harness. Deployment requires neither weight updates nor teacher calls. Across seven visual question-answering settings with students of at most 9B parameters, HC improves scores over bare students by 9.9-23.9 points, averaged over three independent builds per setting. Interventions on five runtime decision types (invocation, selection, argument generation, evidence integration and abstention) show why this allocation matters: requesting evidence and generating open queries can be costly, whereas bounded choices and reading supplied text can remain useful student work. Fact cards benefit all ten evaluated students, but decision policies transfer unevenly. Recompilation for a new student model helps when the transferred interface no longer fits the student. With 100 practice items, HC exceeds answer-only LoRA on three tasks. Larger training budgets can match or surpass a fixed harness, while combining the two improves SlideVQA beyond either alone. These findings support allocating work from measured student behavior rather than uniformly removing decisions. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.11231 — 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 2026-10-09T05:47:00.598Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).