כתבה
arXiv cs.LG ·
בחינה: כל מה שצריך - תכנון אסטרטגי להצגת יכולות התפיסה של LLM
Evaluation is All You Need: Strategic Overclaiming of LLM Reasoning Capabilities Through Evaluation Design
מחקר חשף כי תוצאות הבחינה של LLMs חשופות לשינויים גדולים. המחברים טוענים כי זה קשה לחזור על תוצאות טובות באופן יציב.
תקציר מקורי באנגליתarXiv:2506.04734v3 Announce Type: replace-cross Abstract: Reasoning models represented by the Deepseek-R1-Distill series have been widely adopted by the open-source community due to their strong performance in mathematics, science, programming, and other domains. However, our study reveals that their benchmark evaluation results are subject to significant fluctuations caused by various factors. Subtle differences in evaluation conditions can lead to substantial variations in results. Similar phenomena are observed in other open-source inference models fine-tuned based on the Deepseek-R1-Distill series, as well as in the QwQ-32B model, making their claimed performance improvements difficult to reproduce reliably. Therefore, we advocate for the establishment of a more rigorous paradigm for m
קרא במקור המקורי
arxiv.org
פתח כתבה מקורית