Metraipolis › archive › dc_df06dd5c7d14
FOCUS: Decoupling Expert Personas in LLMs to Enhance Domain Expert Capabilities — arXiv:2608.05611
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.05611
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-25T21:27:10.701Z
- SHA-256 of the archived text
7c3a8ca1adcb0c1198ac971f9287e5aaa9c04127993cb8d95790595228fb7275- Stable URL of this page
https://metraipolis.com/documento/dc_df06dd5c7d14
How to cite
FOCUS: Decoupling Expert Personas in LLMs to Enhance Domain Expert Capabilities — arXiv:2608.05611 — arXiv (en), https://arxiv.org/abs/2608.05611 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_df06dd5c7d14 (Metraipolis archive dc_df06dd5c7d14, sha256 7c3a8ca1adcb0c1198ac971f9287e5aaa9c04127993cb8d95790595228fb7275)
BibTeX
@misc{dcdf06dd5c7d14,
title = {FOCUS: Decoupling Expert Personas in LLMs to Enhance Domain Expert Capabilities — arXiv:2608.05611},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.05611},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_df06dd5c7d14; sha256 7c3a8ca1adcb0c1198ac971f9287e5aaa9c04127993cb8d95790595228fb7275},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.05611. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large Language Models (LLMs) can exhibit diverse personas, and activating expert personas has been shown to improve domain expertise and task accuracy. However, existing persona control methods often suffer from cross-domain coupling, which may lead to overly aggressive behavior in high-caution domains such as healthcare, or excessive conservatism in risk-sensitive domains such as financial trading. To address this issue, we propose FOCUS (\textbf{\underline{F}}ine-tuning with \textbf{\underline{O}}rthogonal \textbf{\underline{C}}ontrol for \textbf{\underline{U}}ncoupled persona\textbf{\underline{S}}). FOCUS first automatically extracts expert persona vectors from LLMs, then applies orthogonal decomposition to decouple domain-specific expert personas, and finally introduces an expert gating module to adaptively control persona activation according to task contexts. With a two-stage training strategy and a gated selection regularizer, the model learns to activate appropriate personas for both single-domain and cross-domain tasks. Experiments on financial, legal, medical, and cross-domain benchmarks show that FOCUS improves task accuracy and outperforms existing persona control methods. Our code is available at \href{ this https URL }{this url}.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-07 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.05611 — 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 2026-10-25T21:24:38.196Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).