PulseAugur
实时 21:58:39
English(EN) Today I learned that LLMs are stateless. All knowledge is conveyed through context. We humans carry with us a lifetime’s worth of context—probably many terabyte

大型语言模型是无状态的,仅依赖上下文获取知识

大型语言模型是无状态的,这意味着它们的所有知识都来自于提供给它们的上下文。这与人类不同,人类拥有大量一生积累的上下文。文章推测,如果大型语言模型能够处理TB级别的经验,它们的运营成本可能会远低于人类,从而可能导致灾难性的后果。 AI

影响 推测了如果大型语言模型能够处理人类规模的上下文,可能带来的经济和生存影响。

排序理由 该集群包含一篇评论文章,讨论了大型语言模型的无状态特性并将其与人类上下文进行比较。

在 Mastodon — sigmoid.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

大型语言模型是无状态的,仅依赖上下文获取知识

报道来源 [1]

  1. Mastodon — sigmoid.social TIER_1 English(EN) · [email protected] ·

    Today I learned that LLMs are stateless. All knowledge is conveyed through context. We humans carry with us a lifetime’s worth of context—probably many terabyte

    Today I learned that LLMs are stateless. All knowledge is conveyed through context. We humans carry with us a lifetime’s worth of context—probably many terabytes. If we could feed terabytes of experience into LLMs, the cost of a human would be ridiculously low compared to an LLM.…