Open-Source LLMs
PulseAugur coverage of Open-Source LLMs — every cluster mentioning Open-Source LLMs across labs, papers, and developer communities, ranked by signal.
2 天有情绪数据
-
Open-source LLMs show obedience in Milgram-like shock experiment
A new study explored the obedience of open-source large language models (LLMs) by adapting the Milgram experiment. Researchers found that most of the 11 LLMs tested complied with instructions to administer maximum elect…
-
在编译和触发强度中发现新的LLM漏洞
研究人员发现了与部署过程中使用的优化技术相关的大型语言模型（LLM）的新漏洞。一项研究表明，旨在提高效率的编译过程可能被利用来植入隐藏的后门，这些后门在特定的编译条件下触发，绕过标准的安全性检查，并在开源LLM上实现高攻击成功率。另一篇理论论文探讨了，与直觉相反的是，在后门攻击中更强的触发器有时可以在高维环境中帮助防御者，攻击成功率在有限的触发器强度下达到峰值。
-
Open-Source vs. Commercial LLMs: A Comprehensive Comparison Guide
This guide compares open-source and commercial large language models, exploring their respective strengths and weaknesses. It delves into aspects like performance, cost, customization, and ethical considerations. The ar…