Metraipolis › archive › dc_b19e277664dd

Model vs. User: Whose Values Prevail? Inference-Time Value Alignment in Multi-Turn Conversations — arXiv:2610.19817

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19817
License (as archived)
CC0-1.0
Archived
2026-10-15T20:34:45.008Z
SHA-256 of the archived text
2fda4e1c5502cb13ad5af7c1f5816df3915ed68c344863e9c37219a368e37baa
Stable URL of this page
https://metraipolis.com/documento/dc_b19e277664dd

How to cite

Model vs. User: Whose Values Prevail? Inference-Time Value Alignment in Multi-Turn Conversations — arXiv:2610.19817 — arXiv (en), https://arxiv.org/abs/2610.19817 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b19e277664dd (Metraipolis archive dc_b19e277664dd, sha256 2fda4e1c5502cb13ad5af7c1f5816df3915ed68c344863e9c37219a368e37baa)
BibTeX
@misc{dcb19e277664dd,
  title = {Model vs. User: Whose Values Prevail? Inference-Time Value Alignment in Multi-Turn Conversations — arXiv:2610.19817},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19817},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b19e277664dd; sha256 2fda4e1c5502cb13ad5af7c1f5816df3915ed68c344863e9c37219a368e37baa},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19817. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Human values shape how we expect language models to behave. Dominant alignment approaches treat values as global objectives, although the appropriate value orientation is personal and contextual. Training-free alignment can steer models toward target values at inference time, yet has been largely evaluated in single-turn settings and with a single steering direction (e.g., toward safety). We evaluate four prompt- and activation-based steering methods in multi-turn conversations, where user and assistant models push toward the same or opposing, equally desirable values. We track alignment with a linear probe across three models and two value trade-offs. When assistant and user pull in opposite directions, the side matching the assistant's baseline bias often prevails, and effects that look successful early often fade or reverse after a few turns. Steering should therefore be evaluated over full conversations and in both directions of a value trade-off.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.19817 — 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 2026-10-15T19:36:18.226Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).