arXiv cs.CL· Yiwen Yang, Guanming Liu·· 11 小时前AI 评分31
ProcessPrefBench：评估自然人机交互中个体过程偏好的跨任务迁移
ProcessPrefBench: Evaluating Cross-Task Transfer of Individual Process Preferences from Naturalistic Human-AI Interaction
AI 导读
ProcessPrefBench 基于两位知识工作者三个月的自然对话构建 66 个源-目标案例，测试模型能否把一次对话中暴露的过程偏好迁移到未重述该偏好的后续任务。相比仅给目标提示，完整历史和指令化蒸馏偏好提示分别将偏好实现度提升 0.204 和 0.244 分（0-2 量表），但蒸馏无可靠优势。捏造证据的失败率同时从 5.8% 升至 10.1% 和 10.4%。
正文 · 原文
Abstract:In natural human-AI dialogue, users' corrections, requests, and self-descriptions carry reusable signals about how they want work to be done. These signals reveal individual process preferences, user-specific standards for evidence, depth, analytical focus, procedure, and delivery. We ask whether a preference revealed in one dialogue can shape an assistant's later answer when the task does not restate it. ProcessPrefBench uses three months of natural dialogues from two knowledge workers to build 66 source-target cases separately judged for source evidence and target applicability. Sources are preselected, so the benchmark tests application rather than retrieval. We test three models with target-only, full-history, and distilled-preference prompts. Target-specific 0-2 rubrics score preference realization, and a task gate flags severe failures. Compared with target-only prompting, full-history and imperative distilled-preference prompting raise preference realization by 0.204 and 0.244 points on the 0-2 scale, respectively, with no reliable advantage from distillation. Fabricated-evidence failures rise from 5.8% under target-only prompting to 10.1% under full-history and 10.4% under imperative prompting. ProcessPrefBench highlights a tension in learning from dialogue, where drawing on prior interactions can improve preference realization yet also invite fabricated evidence. We hope it helps build assistants that honor a user's process preferences without sacrificing factual reliability.
| Comments: | 11 pages, 2 figures. Published at LM Playschool 2026 (EMNLP 2026 workshop) |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.06065 [cs.CL] |
| (or arXiv:2611.06065v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.06065 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | The Proceedings of Improving Language Models through Learning from Dialogue Game Interaction, pages 30-40, Association for Computational Linguistics, 2026 |
| Related DOI: | https://doi.org/10.18653/2026.lmplayschool-1.3
DOI(s) linking to related resources |
Submission history
From: Yiwen Yang [view email]
[v1]
Wed, 4 Nov 2026 07:10:46 UTC (578 KB)
来源：arXiv cs.CL · arxiv.org