An Analysis and Mitigation of the Reversal Curse 的分析与缓解
An Analysis and Mitigation of the Reversal Curse
AI总结：
本文首次分析了LLMs中“反转诅咒”现象的成因，发现其源于以next-token prediction为代表的特定训练目标，呼吁学界关注该问题及大模型的其他潜在局限。
摘要：
近期的研究在大语言模型（LLMs）中观察到了一个显著现象，被称为“reversal curse（反转诅咒）”。反转诅咒是指，当处理由关系 $R$ 及其逆关系 $R^{-1}$ 连接的两个实体 $a$ 和 $b$ 时，LLMs 擅长处理“$aRb$”形式的序列，但在处理“$bR^{-1}a$”形式的序列时，无论是在生成还是理解方面都会遇到挑战。例如，GPT-4 能够准确地回答“Tom Cruise 的母亲是？”这一问题，回答“Mary Lee Pfeiffer”，但当被问及“Mary Lee Pfeiffer 的儿子是？”时，却难以给出令人满意的答案。在本文中，我们首次研究了反转诅咒在 LLMs 中是如何发生的。我们的调查表明，反转诅咒可能源于特定的训练目标，这在大多数因果语言模型中广泛使用的 next-token prediction（下一个 token 预测）中尤为明显。我们希望这项初步研究能引起人们对反转诅咒以及当前 LLMs 中其他潜在局限性的更多关注。
英文摘要：
Recent research observed a noteworthy phenomenon in large language models (LLMs), referred to as the ``reversal curse.'' The reversal curse is that when dealing with two entities, denoted as $a$ and $b$, connected by their relation $R$ and its inverse $R^{-1}$, LLMs excel in handling sequences in the form of ``$aRb$,'' but encounter challenges when processing ``$bR^{-1}a$,'' whether in generation or comprehension. For instance, GPT-4 can accurately respond to the query ``Tom Cruise's mother is?'' with ``Mary Lee Pfeiffer,'' but it struggles to provide a satisfactory answer when asked ``Mary Lee Pfeiffer's son is?'' In this paper, we undertake the first-ever study of how the reversal curse happens in LLMs. Our investigations reveal that the reversal curse can stem from the specific training objectives, which become particularly evident in the widespread use of next-token prediction within most causal language models. We hope this initial investigation can draw more attention to the reversal curse, as well as other underlying limitations in current LLMs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
