让大模型别太聪明，才能更好模拟人类语言预测
To model human linguistic prediction, make LLMs less superhuman
- 让大模型记忆能力接近人类，避免过度超人化
- 当前大模型预测能力越强，越难解释人类阅读行为
- 适合研究语言认知与人机对齐的学者参考
阅读时，人类会预测下一个词，这一过程影响阅读行为。大语言模型（LLMs）因同样具备词序预测能力，被用作人类语言预测的模型。然而近来发现，随着LLMs预测能力提升，其解释人类阅读行为的效果反而下降。我们指出，这是因为当前LLMs的预测能力远超人类：这由海量训练数据、更强的长期记忆和短期记忆驱动。我们主张构建具有类人记忆的大模型，并设计新实验衡量人类与模型间的对齐度，提出实现路径。
原文摘要 · Abstract (English)
When we read, we make predictions about upcoming words; these predictions influence our reading behavior. The success of large language models (LLMs), which, like humans, make predictions about upcoming words, has motivated their use as models of human linguistic prediction. Surprisingly, in the last few years, as LLMs' ability to predict the next word has improved, their ability to explain reading behavior has declined. We argue this is because current LLMs can predict upcoming words much better than human readers can. This 'superhumanness' is driven by LLMs' extensive training data, stronger long-term memory of training examples, and stronger short-term memory. We advocate for LLMs with human-like memory and for new experiments to measure the alignment between humans and LLMs, and outline directions towards achieving these goals.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。