Zhizhuo Kou, Heming Li, Jian Yang, Junyu Luo, Jingshu Peng, Sirui Han, Yike Guo · 2026 · arXiv
DOI: 10.48550/arXiv.2610.28474 · open-access PDF
그러나 채팅 모델은 사용자가 사용하는 방언에 관계없이 표준 영어로 응답하는 경향이 있습니다. 일관된 대답은 교육이 이러한 사용자를 동일하게 대우하는지 여부를 숨길 수 있으며, 출력 감사도 말할 수 없습니다. 우리는 감독된 미세 조정 (SFT)에서 선호 최적화 (DPO) 및 강화 학습에 이르기까지 각 단계가 사용자에게 맞는 모델의 경향을 변화시키는 방법을 추적합니다.
Machine translation (Meta M2M100); the English original follows.
People adapt how they talk to the person in front of them. However, chat models tend to respond in Standard English regardless of the dialect used by the user. A uniform answer can hide whether training treats these users alike, and output audits cannot tell as well. We trace how each post-training stage, from supervised fine-tuning (SFT) to preference optimization (DPO) and reinforcement learning, changes a model's tendency to match the user, and read what each update prefers from two content-matched replies, one matching the user and one not. Across 10 open model lineages, post-training stan
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.28474, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange