Metraipolis › archive › dc_1042cb0eae2e
OSPD: On-Policy Self-Distillation for Persona-Consistent Dialogue — arXiv:2609.34418
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34418
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:57:45.721Z
- SHA-256 of the archived text
c9dd47bee1cdb5313d3b1dc56b2c53addd7ab76135107b46ebc785664b44a359- Stable URL of this page
https://metraipolis.com/documento/dc_1042cb0eae2e
How to cite
OSPD: On-Policy Self-Distillation for Persona-Consistent Dialogue — arXiv:2609.34418 — arXiv (en), https://arxiv.org/abs/2609.34418 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1042cb0eae2e (Metraipolis archive dc_1042cb0eae2e, sha256 c9dd47bee1cdb5313d3b1dc56b2c53addd7ab76135107b46ebc785664b44a359)
BibTeX
@misc{dc1042cb0eae2e,
title = {OSPD: On-Policy Self-Distillation for Persona-Consistent Dialogue — arXiv:2609.34418},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34418},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1042cb0eae2e; sha256 c9dd47bee1cdb5313d3b1dc56b2c53addd7ab76135107b46ebc785664b44a359},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34418. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Maintaining persona consistency across multi-turn dialogues remains a core challenge for role-playing language models. Off-policy distillation from external teachers incurs distribution mismatch that compounds across dialogue turns, while reinforcement learning struggles with reward ambiguity inherent in subjective persona fidelity. We propose OSPD, an on-policy self-distillation framework where the same model serves as both teacher and student under asymmetric information: the teacher receives a complete character profile while the student sees only a brief summary, and the student generates trajectories from its own policy. We find that teacher confidence in role-playing dialogue exhibits a bimodal structure---sharply peaked at character-critical tokens yet diffuse at generic utterances---and introduce role-aware divergence switching to match this structure. A progressive trait masking curriculum further forces staged internalization of character knowledge along semantic dimensions. Experiments on CharacterBench, CharacterEval, and SocialBench show that OSPD substantially improves persona consistency over supervised fine-tuning and multi-turn RL baselines, without requiring any external teacher or reward model. 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/2609.34418 — 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 2026-10-07T20:29:01.546Z 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).