引入了一个新的基准测试 StreamProfileBench，用于评估大型语言模型 (LLM) 从连续到达的内容中推断用户画像的能力，这项任务是当前静态评估所忽略的。该基准测试包含一个数据集，其中包含来自 7,000 多名真实用户在五个平台上的 120,000 多条用户生成内容帖子。对 14 个 LLM 的实验显示，模型在连续更新用户画像方面存在困难，表现出保留旧兴趣的偏见，并且未能识别衰退的兴趣。 AI
影响 凸显了 LLM 在实时系统中适应不断变化的用��兴趣方面的局限性。
排序理由 该集群描述了一篇介绍用于评估 LLM 的基准测试和数据集的新学术论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.25758v1 Announce Type: new Abstract: Large Language Models (LLMs) have reshaped user profiling, yet current evaluations mainly focus on static data snapshots. This paradigm overlooks the reality of personalized systems, where User-Generated Content (UGC) arrives contin…
Large Language Models (LLMs) have reshaped user profiling, yet current evaluations mainly focus on static data snapshots. This paradigm overlooks the reality of personalized systems, where User-Generated Content (UGC) arrives continuously and fine-grained profile evolve rapidly. …
Large Language Models (LLMs) have reshaped user profiling, yet current evaluations mainly focus on static data snapshots. This paradigm overlooks the reality of personalized systems, where User-Generated Content (UGC) arrives continuously and fine-grained profile evolve rapidly. …