📄 中文摘要
大型语言模型（LLM）的快速发展导致基准测试文献面临严重的过时危机。随着科技公司不断更新其专有LLM，新的版本频繁推出，旧版本被弃用，造成基准测试文献引用了过时的模型。这种快速迭代和部署的机制使得学术研究与行业创新之间存在系统性的错位，影响了模型评估的时效性和相关性。为了解决这一问题，提出了多种策略，以确保基准测试能够及时反映最新的技术进展，保持其在研究和应用中的有效性。
应对大型语言模型基准测试过时危机：及时和相关模型评估的策略
📄 English Summary
Addressing LLM Benchmarking Obsolescence: Strategies for Timely and Relevant Model Evaluation
The rapid advancement of Large Language Models (LLMs) has led to a significant obsolescence crisis in benchmarking literature. Tech companies continuously update their proprietary LLMs, frequently releasing new versions while deprecating older ones, which results in benchmarking papers referencing outdated models. This mechanism of rapid iteration and deployment creates a systemic mismatch between academic research and industry innovation, affecting the timeliness and relevance of model evaluations. Various strategies are proposed to address this issue, ensuring that benchmarking remains reflective of the latest technological advancements and maintains its effectiveness in both research and application.