Metraipolis › archive › dc_44c975391d97
Fewer Tokens, More Answers: Verifiable-Reward Tuning of an Ornith-Style VLM for Multi-View Driving QA — arXiv:2611.00833
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.00833
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T18:24:34.186Z
- SHA-256 of the archived text
8c7dfcb78d380d97a9e571cf89db62bcf95acd9b5c251fb7f55a6f81ac0372d1- Stable URL of this page
https://metraipolis.com/documento/dc_44c975391d97
How to cite
Fewer Tokens, More Answers: Verifiable-Reward Tuning of an Ornith-Style VLM for Multi-View Driving QA — arXiv:2611.00833 — arXiv (en), https://arxiv.org/abs/2611.00833 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_44c975391d97 (Metraipolis archive dc_44c975391d97, sha256 8c7dfcb78d380d97a9e571cf89db62bcf95acd9b5c251fb7f55a6f81ac0372d1)
BibTeX
@misc{dc44c975391d97,
title = {Fewer Tokens, More Answers: Verifiable-Reward Tuning of an Ornith-Style VLM for Multi-View Driving QA — arXiv:2611.00833},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.00833},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_44c975391d97; sha256 8c7dfcb78d380d97a9e571cf89db62bcf95acd9b5c251fb7f55a6f81ac0372d1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00833. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reasoning-enabled model derivatives must turn intermediate deliberation into a usable answer within a deployment budget. We study that transition in an Ornith checkpoint adapted to multi-camera driving question answering with verifiable rewards. Paired archived outputs show that, on a selected 400-question development screen with a 256-token cap, GRPO increases answer extraction from 78.25% to 90.25% and accuracy from 55.50% to 63.75%. Of 39 wrong-to-correct transitions, 33 begin with an unextractable base response. The base's 87 unextractable responses all reach the token cap. Checkpoint-verified reconstruction on those same questions gives a much smaller first-pass accuracy gain at 512 tokens: +0.75 percentage points, versus +8.25 at 256. A later answer-cued continuation is analyzed separately; it helps the base more and narrows the adapter difference. On the full 2,511-question development set, this recovery-enabled pipeline reduces mean generated tokens from 108.6 to 47.1, while the paired accuracy interval includes zero gain. We frame these findings as adaptation of answer delivery under finite budgets, with explicit selection and provenance limits. Ornith's self-scaffolding lineage and newer reasoning-enabled derivatives such as Bonsai motivate the problem, but our measured results concern Ornith alone and do not establish a shared architectural mechanism or cross-model transfer. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Computation and Language (cs.CL); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.00833 — 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 2026-11-03T18:10:08.022Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).