כתבה
arXiv cs.AI ·
הקשר המוטעה: תפיסה של תבניות כאילו-טבעיות במודלי שפה גדולים
Illusory Pattern Perception Drives Spurious Inference in Large Language Models
מודלי שפה גדולים חושבים תבניות כאילו-טבעיות, וזה גורם לשגיאות מערכתיות. קוד זמין ב-https://github.com/NusIoraPrivacy/illusory.
תקציר מקורי באנגליתarXiv:2610.07791v1 Announce Type: new Abstract: Illusory pattern perception is a well-documented human cognitive tendency to infer meaningful relationships in data that is actually random. Such a tendency, often described as "connecting the dots" where none exist, can result in systematic reasoning errors. This paper investigates whether Large Language Models (LLMs) exhibit such perceptual tendencies, which can lead to systematic errors in downstream applications. To our knowledge, this work presents the first systematic study of illusory pattern perception in LLMs, adapting classic psychological paradigms to three tasks with direct empirical comparison to human behaviors. We find that LLMs frequently exhibit stronger illusory pattern perception than humans. In particular, models tend to o
קרא במקור המקורי
arxiv.org
פתח כתבה מקורית