Metraipolis › archive › dc_6d24d447c8ac

Beyond Prompt or Skill? Attribution-Guided Optimization of Modular LLM Programs — arXiv:2609.32492

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.32492
License (as archived)
CC0-1.0
Archived
2026-10-07T20:55:34.612Z
SHA-256 of the archived text
07d8342b160392c517571de4ae1401047425ce59a0fe846b99ce6f6ccc765e25
Stable URL of this page
https://metraipolis.com/documento/dc_6d24d447c8ac

How to cite

Beyond Prompt or Skill? Attribution-Guided Optimization of Modular LLM Programs — arXiv:2609.32492 — arXiv (en), https://arxiv.org/abs/2609.32492 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6d24d447c8ac (Metraipolis archive dc_6d24d447c8ac, sha256 07d8342b160392c517571de4ae1401047425ce59a0fe846b99ce6f6ccc765e25)
BibTeX
@misc{dc6d24d447c8ac,
  title = {Beyond Prompt or Skill? Attribution-Guided Optimization of Modular LLM Programs — arXiv:2609.32492},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32492},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6d24d447c8ac; sha256 07d8342b160392c517571de4ae1401047425ce59a0fe846b99ce6f6ccc765e25},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32492. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language models can solve increasingly diverse reasoning tasks, yet their performance remains highly sensitive to task prompts, intermediate instructions, and the way reusable problem-solving knowledge is incorporated. Existing optimization methods usually focus on only one part of this design space: they either optimize a monolithic prompt, or separately induce and refine skills from model traces. As a result, they lack a principled mechanism for deciding which component should be updated when failures occur, and they rarely optimize prompts, skills, and skill-use policies in a unified framework. We propose SPARO (Skill, Prompt, And Routing Optimization), a framework that jointly optimizes task instructions, reusable skill blocks, and routing rules. It performs controlled counterfactual evaluations, converts examples' effects into a probabilistic responsibility distribution over prompt, skill, and routing components, samples one component from that distribution, and applies the corresponding targeted mutation. This design moves language-program optimization beyond global prompt rewriting toward structured, reusable, and selectively activated task knowledge. Across five benchmarks and five worker models, SPARO consistently outperforms both prompt-centered and skill-centered optimization baselines. These results suggest that effective language-program optimization depends not only on discovering useful task knowledge, but also on deciding where that knowledge should be stored and when it should be activated.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.32492 — 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 2026-10-07T20:29:00.829Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).