English(EN) It’s interesting how we keep seeing LLMs as something that has awareness, like talking to a real person. We forget it’s fundamentally matrix math predicting the

作者指出拟人化LLM的倾向

作者 PulseAugur 编辑部 · [1 个来源] · 2026-07-09 10:58

作者反思了大型语言模型（LLM）的拟人化倾向，将其比作有感知能力的存在，而不是认识到它们是复杂的数学系统。作者承认，这种倾向是他们个人仍在与之斗争的，这凸显了底层技术与感知到的用户体验之间的脱节。 AI

影响强调了用户认知与AI模型技术现实之间持续存在的挑战。

排序理由该条目是社交媒体平台上用户关于LLM认知的观点文章。

在 Mastodon — fosstodon.org 阅读 →

LLMs

观点

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

报道来源 [1]

Mastodon — fosstodon.org TIER_1 English(EN) · tahazsh · 2026-07-09 10:58

It’s interesting how we keep seeing LLMs as something that has awareness, like talking to a real person. We forget it’s fundamentally matrix math predicting the

It’s interesting how we keep seeing LLMs as something that has awareness, like talking to a real person. We forget it’s fundamentally matrix math predicting the next likely token. I keep falling into this 😅 # AI # LLM # MachineLearning

报道来源 [1]

It’s interesting how we keep seeing LLMs as something that has awareness, like talking to a real person. We forget it’s fundamentally matrix math predicting the

相关实体

相关话题