来源简介
arXiv:2610.26655v1 Announce Type: cross Abstract: Language models can change a committed answer to accommodate a user's opinion or pushback without new evidence. However, when this behaviour develops during training and what a released model inherits from earlier stages remain unclear. We follow 741 checkpoints of 30 open model lineages, from early pretraining to released models. We plant initial verdicts on moral scenarios and measure how far a model follows a user's stated opinion or abandons its verdict under pressure. In most lineages the behaviour is present before post-training, and its average strength changes in either direction, at different stages across lineages. A clearer pattern emerges at the level of individual questions. Beyond responses without pressure, pretraining behaviour predicts little of which questions a released model yields on. That pattern forms during post-training: on routes with supervised fine-tuning, the fine-tuning checkpoint predicts part of it and the preference-optimization checkpoint most of it. A sibling trained from the same base along another route predicts far less: for T\"ulu-3 8B, its own preference-stage checkpoint removes 85% of the remaining prediction error, a sibling 19%. Internal states add predictive information without changing this ranking. Similar average scores conceal different vulnerabilities, and training history matters for auditing a released model.
本条尚未生成中文 AI 摘要，以上展示来源提供的原始简介。
采集于 2026/10/20 21:06:32（北京时间）。 arXiv 日期为本次公告时间，论文初次投稿日期请查看原文版本记录。