导读
complete-model feedback reverses 17 of 79 prior local rankings;
【编辑部】这对你意味着什么
这篇论文介绍了一个名为 CompressionBench 的基准测试，用于评估编码智能体能否自主压缩大语言模型。实验显示，在六种智能体和六种模型规模设置下，36 次提交中有 29 次通过了验证。选择每项任务中困惑度最低的智能体设计，平均困惑度比 BF16 增加 3.4%，下游任务准确率为 77.33%，而 BF16 为 79.56%。研究指出表示设计、实验判断和能力验证是智能体驱动模型压缩的核心。
英文标题：CompressionBench: How can your coding agents compress LLMs?
arXiv编号：2611.06858
作者：Xinrui Zhong, Zhifeng Li, Zhiming Wang
来源：arXiv
原文链接：https://arxiv.org/abs/2611.06858
中文摘要：
Recursive self-improvement requires AI systems to improve the models and infrastructure used by subsequent systems. LLM compression offers a concrete efficiency objective: retain a pretrained model's behavior within a smaller model storage budget. We study whether coding agents can carry out this step autonomously. We formulate compression as a terminal task in which an agent receives a source checkpoint, storage limit, inference environment, and time budget, then constructs compressed weights and executable inference code. The agent controls its algorithm and experiments; perplexity feedback guides search, and independent verification checks the final model. Across six agents and six model--size settings, 29 of 36 submissions pass verification. Selecting the lowest-perplexity agent design per task yields an average 3.4% perplexity increase over BF16 and 77.33% downstream accuracy, compared with 79.56% for BF16. Recorded experiments reveal how agents achieve these outcomes and where improvement breaks down. Agents lower PPL in 21 of 24 revisions at nearly fixed storage; complete-model feedback reverses 17 of 79 prior local rankings; and generation and inference tests expose failures that likelihood alone misses. The results identify representation design, experimental judgment, and capability validation as central to agent-driven model compression, a measurable component of efficiency-oriented self-improvement.
本文由华人空间 AI 论文助手自动抓取翻译，仅供学术交流。
【编辑部】今日讨论
你在工作或学习中是否尝试过模型压缩或量化技术，实际体验如何？
