跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.25594 · v1

Bradley-Terry Training Can Distort Outcome Probabilities of Initially Accurate Language World Models

首次提交 2026-11-19 14:42
版本更新 2026-11-19 14:42
雷达观测 2026-11-20 19:58
arXiv ↗PDF ↗

摘要

Comparative feedback, in which a judge says which of two candidate outcomes is closer to what happened, is a practical way to train language world models that assign probabilities to an action's outcomes. We show that, fitted with a Bradley-Terry objective, it need not teach the true outcome probabilities and can distort accurate ones. On a simulated robot push with enumerated outcome distributions, this objective takes a model that is accurate on unseen layouts (hazard error 2.4 percentage points, pp) and moves 33 to 43 pp of its probability onto outcomes that none of a condition's enumerated branches produces. The cause is the objective, not the comparisons: its optimum, computed before training, predicts where the model ends up, and a likelihood that knows the judge's rule learns from the same comparisons without the shift. The shift is robust: it recurs on unseen conditions, under five candidate laws and a three-fold denser enumeration, and about 29 pp remain without the physically impossible events. Its cost is conditional: with those events it inflates hazard probabilities, and a threshold inherited from the pre-feedback model costs about 3 pp of utility per decision on trained and 1 to 2 pp on unseen conditions, through false rejects; without them the utility loss averages below 1 pp over six seeds, though a small regret in action choice remains. Audits of readout, response and decision trace the effect from probabilities to actions.

作者备注:10 pages, 6 figures, 1 table

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-19 14:42 · 最新观测版本

Bradley-Terry Training Can Distort Outcome Probabilities of Initially Accurate Language World Models

Comparative feedback, in which a judge says which of two candidate outcomes is closer to what happened, is a practical way to train language world models that assign probabilities to an action's outcomes. We show that, fitted with a Bradley-Terry objective, it need not teach the true outcome probabilities and can distort accurate ones. On a simulated robot push with enumerated outcome distributions, this objective takes a model that is accurate on unseen layouts (hazard error 2.4 percentage points, pp) and moves 33 to 43 pp of its probability onto outcomes that none of a condition's enumerated branches produces. The cause is the objective, not the comparisons: its optimum, computed before training, predicts where the model ends up, and a likelihood that knows the judge's rule learns from the same comparisons without the shift. The shift is robust: it recurs on unseen conditions, under five candidate laws and a three-fold denser enumeration, and about 29 pp remain without the physically impossible events. Its cost is conditional: with those events it inflates hazard probabilities, and a threshold inherited from the pre-feedback model costs about 3 pp of utility per decision on trained and 1 to 2 pp on unseen conditions, through false rejects; without them the utility loss averages below 1 pp over six seeds, though a small regret in action choice remains. Audits of readout, response and decision trace the effect from probabilities to actions.

查看此版本解读 →查看此版本原文 ↗