Metraipolis › archive › dc_59c8a2bb6a6f
When a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning Models — arXiv:2609.39846
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.39846
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T12:45:50.341Z
- SHA-256 of the archived text
6bd11a6d482f71f065e1ec8401a71f0f924eed6c7245009b34942a5c665daebc- Stable URL of this page
https://metraipolis.com/documento/dc_59c8a2bb6a6f
How to cite
When a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning Models — arXiv:2609.39846 — arXiv (en), https://arxiv.org/abs/2609.39846 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_59c8a2bb6a6f (Metraipolis archive dc_59c8a2bb6a6f, sha256 6bd11a6d482f71f065e1ec8401a71f0f924eed6c7245009b34942a5c665daebc)
BibTeX
@misc{dc59c8a2bb6a6f,
title = {When a Kindergartener Solves Calculus: Measuring Capability Leakage in Role-Prompted Reasoning Models — arXiv:2609.39846},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.39846},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_59c8a2bb6a6f; sha256 6bd11a6d482f71f065e1ec8401a71f0f924eed6c7245009b34942a5c665daebc},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39846. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We investigate the problem of role-capability leakage (RCL), in which a role-prompted reasoning model generates convincing in-role text while continuing to exhibit capabilities on benchmarks that exceed those implied by the assigned role. For example, when a model is prompted to assume the role of a kindergarten student, one might expect its performance on a mathematics benchmark to reflect kindergarten-level ability rather than expert-level proficiency in solving calculus problems. We introduce RoleCapBench, a curriculum-grounded benchmark for evaluating RCL across six educational roles and four assessment levels spanning elementary school through A-level, and use it to evaluate three open-weight reasoning models. We find that although the models can generate stylistically convincing in-role responses, they consistently fail to align their underlying capabilities with their assigned roles. Naive role prompting yields strong role-voice scores of 1.218--1.389 while retaining above-role accuracy of 0.811--0.898. RCL persists across a range of prompting conditions, including prompts that explicitly instruct the model to match the role's capability level. To mitigate this problem, we propose Injection, an inference-time intervention that combines explicit, role-specific capability guidelines with a guiding prefilled response prefix. Injection improves role-capability alignment across models, reducing above-role accuracy by up to 0.562 while preserving in-role accuracy with a marginal drop of less than 0.058 across most models. All artifacts, including scripts and evaluation data, will be released upon acceptance. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.AI). Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC). Pagina del lavoro: https://arxiv.org/abs/2609.39846 — 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 2026-10-07T12:14:14.161Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).