arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.14188cs.CLcs.AIcs.HC
收藏

GPT-5与其他LLM在长短上下文性能中的比较

GPT-5 vs Other LLMs in Long Short-Context Performance

Bernoulli Institute RUG(伯努利研究所);ISRC, Khazar University(ISRC、哈扎尔大学);Department of Computer Engineering University of Guilan(计算机工程系、盖兰大学);QCE Department TU Delft(QCE部门、代尔夫特理工大学)

Nima Esmi, Maryam Nezhad-Moghaddam, Fatemeh Borhani, Asadollah Shahbahrami, Amin Daemdoost, Georgi Gaydadjiev

AI总结:

本文比较了GPT-5与其他LLM在长短上下文任务中的性能,发现GPT-5在高输入量下保持高精度,揭示了模型在复杂任务中的实际表现与理论能力之间的差距。

摘要:

随着大型语言模型(LLM)上下文窗口的显著扩展,这些模型理论上能够单次处理数百万个标记。然而,研究表明,这种理论能力与模型实际稳健利用长上下文信息的能力之间存在显著差距,尤其是在需要全面理解众多细节的任务中。本文评估了四种最先进的模型(Grok-4、GPT-4、Gemini 2.5和GPT-5)在长短上下文任务上的性能。为此,使用了三个数据集:两个补充数据集用于检索食谱和数学问题,以及一个包含20,000条社交媒体帖子的主要数据集用于抑郁症检测。结果显示,当社交媒体数据集的输入量超过5,000条帖子(70,000个标记)时,所有模型的性能显著下降,准确率降至约50-53%。值得注意的是,在GPT-5模型中,尽管准确性急剧下降,其精度仍保持在约95%的高水平,这一特性可能对敏感应用如抑郁症检测具有重要价值。本研究还表明,

英文摘要:

With the significant expansion of the context window in Large Language Models (LLMs), these models are theoretically capable of processing millions of tokens in a single pass. However, research indicates a significant gap between this theoretical capacity and the practical ability of models to robustly utilize information within long contexts, especially in tasks that require a comprehensive understanding of numerous details. This paper evaluates the performance of four state-of-the-art models (Grok-4, GPT-4, Gemini 2.5, and GPT-5) on long short-context tasks. For this purpose, three datasets were used: two supplementary datasets for retrieving culinary recipes and math problems, and a primary dataset of 20K social media posts for depression detection. The results show that as the input volume on the social media dataset exceeds 5K posts (70K tokens), the performance of all models degrades significantly, with accuracy dropping to around 50-53% for 20K posts. Notably, in the GPT-5 model, despite the sharp decline in accuracy, its precision remained high at approximately 95%, a feature that could be highly effective for sensitive applications like depression detection. This research also indicates that the "lost in the middle" problem has been largely resolved in newer models. This study emphasizes the gap between the theoretical capacity and the actual performance of models on complex, high-volume data tasks and highlights the importance of metrics beyond simple accuracy for practical applications.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑