← 研踪 PaperTrailarXiv:2604.17283 · v2
HorizonBench: Long-Horizon Personalization with Evolving Preferences
首次提交 2026-04-19 14:55
版本更新 2026-11-18 06:59
雷达观测 2026-11-19 19:08
摘要
Personal assistants now talk with the same user for months, and users' preferences change as their lives change; an assistant that keeps serving a preference the user has outgrown fails that user. Real conversations do not record when or why a preference changed, and existing benchmarks either hold preferences fixed or have the user state the change, so they cannot tell whether a model forgot what the user said or failed to update its belief about the user. We introduce HorizonBench and its data generator, which samples how a user's preferences change and generates six months of conversation from that sequence; each change follows a life event that the user mentions without stating the new preference. A human longitudinal preference dataset that we collect and release, 83,024 ratings from 160 people over 13 weeks, sets how often and how much life events change preferences. HorizonBench contains 4,245 questions from 360 users, and when a preference has changed, its earlier value is one of the answer options. All 25 frontier models we evaluate choose the outdated preference more often than evenly spread errors would predict, and the best reaches 52.8% accuracy. Two of the newest models score no higher when given only the relevant conversations than with the full history, yet one added sentence stating the change raises their accuracy by 11 to 22 points, and retrieval and memory methods do not recover the lost accuracy. Remembering what a user said and knowing what the user prefers now are therefore separate capabilities, and HorizonBench, its generator, and the human dataset support research on both.
作者备注:26 pages, 6 figures, 19 tables
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。
v2 · 2026-11-18 06:59 · 最新观测版本
HorizonBench: Long-Horizon Personalization with Evolving Preferences
Personal assistants now talk with the same user for months, and users' preferences change as their lives change; an assistant that keeps serving a preference the user has outgrown fails that user. Real conversations do not record when or why a preference changed, and existing benchmarks either hold preferences fixed or have the user state the change, so they cannot tell whether a model forgot what the user said or failed to update its belief about the user. We introduce HorizonBench and its data generator, which samples how a user's preferences change and generates six months of conversation from that sequence; each change follows a life event that the user mentions without stating the new preference. A human longitudinal preference dataset that we collect and release, 83,024 ratings from 160 people over 13 weeks, sets how often and how much life events change preferences. HorizonBench contains 4,245 questions from 360 users, and when a preference has changed, its earlier value is one of the answer options. All 25 frontier models we evaluate choose the outdated preference more often than evenly spread errors would predict, and the best reaches 52.8% accuracy. Two of the newest models score no higher when given only the relevant conversations than with the full history, yet one added sentence stating the change raises their accuracy by 11 to 22 points, and retrieval and memory methods do not recover the lost accuracy. Remembering what a user said and knowing what the user prefers now are therefore separate capabilities, and HorizonBench, its generator, and the human dataset support research on both.
查看此版本解读 →查看此版本原文 ↗