Metraipolis › archive › dc_cf4d40417203

When Users Change Their Minds: Measuring and Repairing Intent Drift in LLM Agents — arXiv:2609.32520

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.32520
License (as archived)
CC0-1.0
Archived
2026-10-07T21:00:19.947Z
SHA-256 of the archived text
4eb8327d1ee4a87b7e921feaf5514cd22485fdd5cee0e908bbf29d8f083b776a
Stable URL of this page
https://metraipolis.com/documento/dc_cf4d40417203

How to cite

When Users Change Their Minds: Measuring and Repairing Intent Drift in LLM Agents — arXiv:2609.32520 — arXiv (en), https://arxiv.org/abs/2609.32520 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cf4d40417203 (Metraipolis archive dc_cf4d40417203, sha256 4eb8327d1ee4a87b7e921feaf5514cd22485fdd5cee0e908bbf29d8f083b776a)
BibTeX
@misc{dccf4d40417203,
  title = {When Users Change Their Minds: Measuring and Repairing Intent Drift in LLM Agents — arXiv:2609.32520},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32520},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cf4d40417203; sha256 4eb8327d1ee4a87b7e921feaf5514cd22485fdd5cee0e908bbf29d8f083b776a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32520. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM agents often operate over multi-turn interactions in which user intent changes before execution. We study intent drift: the failure mode in which superseded parts of the user's intent continue to influence the final answer or tool action. We introduce IntentFlux, an executable benchmark that converts verifiable tasks into dialogues with controlled intent changes while preserving their original graders. In a 627-case calibration, mean task score falls from 0.476 to 0.384 as dialogues contain more superseded and withdrawn information. Across eight models, the rate of fully correct solutions is significantly lower when the same final task must be recovered from an evolving dialogue rather than given directly in a single turn. We further introduce StateForge, which explicitly maintains the active requirements before generation. On General-Test, it improves mean task score from 0.367 to 0.467. Providing the ground-truth final state improves performance further but still does not recover single-turn performance, indicating that state-estimation errors explain only part of the gap. These results establish intent drift as a measurable multi-turn failure mode and explicit state maintenance as a partial mitigation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.32520 — 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 2026-10-07T20:29:02.367Z 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).