《自然》杂志最近发表的一篇论文探讨了大型语言模型（LLMs）是否表现出与拥有心智理论一致的行为。研究表明，LLMs 确实表现出这些特征。作者期待未来的研究能够深入探讨这些能力的潜在机制，特别是使用开源模型。 AI
影响 表明在理解大型语言模型认知能力方面取得了进展，并促使对它们的内部工作机制进行进一步研究。
排序理由 该集群讨论了一篇关于大型语言模型能力的已发表的学术论文。
在 Mastodon — mastodon.social 阅读 →
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
Been enjoying this paper (h/t @[email protected] ) that investigates as well as one can whether LLMs exhibit behavior consistent with possessing a theory of mind. (Spoiler: they do.) Looking forward to a sequel done with open models that opens the black box and te…
Been enjoying this paper (h/t @[email protected] ) that investigates as well as one can whether LLMs exhibit behavior consistent with possessing a theory of mind. (Spoiler: they do.) Looking forward to a sequel done with open models that opens the black box and te…