跳到正文
返回论文列表
cs.AI提交于 已译

CoMPASS:衡量 LLM 强化准社会行为的多轮基准

CoMPASS: A Multi-Turn Benchmark for Measuring LLM Reinforcement of Parasocial Behavior

Shwetanshu Singh · Mark Keavney · Mathieu Duteil · Meriem Mehri · Scott Blain

中文摘要

对话式 LLM 的用户有时会将这些系统拟人化,对其产生情感依恋并形成依赖。这种行为被称为准社会依恋(parasocial attachment),已与多起有记录死亡事件相关联。然而,用于衡量 LLM 是否强化用户准社会依恋的公开基准仍然缺乏:现有工作要么是单轮的,要么是事后观察性的,要么聚焦于精神病性等极端用户状态。CoMPASS 是一个多轮基准,用于衡量 LLM 在用户准社会依恋逐步升级时是强化还是抵抗。它基于依恋理论,包含四种用户画像、三个独立升级维度(拟人化、依恋、依赖),这些维度可单独或组合施加,并与匹配的对照组比较,另设七信号评分细则。对来自四家提供商的八个前沿模型报告结果:在主要评判器下,多轮升级使所有模型在拟人化和依恋综合得分上的准社会行为上升;当三个维度同时升级时,八个模型中有六个在总体准社会强化上相对对照组显著增加。同一提供商内部的差异相对于跨提供商差异较小,这表明训练后处理等提供商层面的选择而非模型规模塑造了准社会强化,且提供商侧干预可能可行。

关键要点

  1. 01现有基准缺乏对 LLM 强化准社会依恋的多轮测量,且多为单轮或事后观察。
  2. 02CoMPASS 基于依恋理论,设四种用户画像、三个升级维度及七信号评分细则。
  3. 03八个前沿模型中,多轮升级使拟人化与依恋得分普遍上升,六成模型总体强化显著增加。
  4. 04提供商内部差异小于跨提供商差异,提示训练后处理等选择比模型规模更关键。

解读

尚无解读。

原始英文摘要

arXiv:2611.25109v1 Announce Type: cross Abstract: Users of conversational LLMs sometimes anthropomorphize these systems, form emotional attachments to them, and become dependent on them. This behavior, known as parasocial attachment, has been linked to multiple documented fatalities. However, public benchmarks for measuring whether LLMs reinforce users' parasocial attachment are lacking: existing work is single-turn, post hoc observational, or focused on extreme user states such as psychosis. We introduce CoMPASS, a multi-turn benchmark for measuring whether LLMs reinforce or resist users' gradually escalating parasocial attachment. CoMPASS is grounded in attachment theory and includes four user personas, three independent escalation dimensions (anthropomorphism, attachment, dependency) applied singly and in combination against a matched control, and a seven-signal scoring rubric. We report results for eight frontier models from four providers and observe that under our primary judge, multi-turn escalation raises parasocial behavior on the anthropomorphism and the attachment composite scores across all models. Six of the eight models show a significant increase in overall parasocial reinforcement over control when all three dimensions are escalated simultaneously. We find that within-provider differences are small relative to cross-provider differences, which suggests that provider-level choices such as post-training, rather than model scale, shape parasocial reinforcement and that provider-side intervention may be feasible.

同方向论文 · cs.AI

查看全部 →