大语言模型的思维链在贝叶斯推理中是否存在幻觉、认知偏差或恐惧症？
Do Chains-of-Thoughts of Large Language Models Suffer from Hallucinations, Cognitive Biases, or Phobias in Bayesian Reasoning?
AI总结：
该研究探究大语言模型思维链在贝叶斯推理中的认知问题，发现其不会主动使用具身启发式等生态效度策略，加入引导提示后采纳仍不稳定，存在符号推理偏差与对生态策略的回避现象，可为贝叶斯推理学习与模型优化提供参考。
摘要：
学会推理并细致解释论证过程，对学生的认知、数学与计算思维发展至关重要。在不确定性问题及贝叶斯推理场景中，这一任务尤其具有挑战性。新一代大语言模型（LLM）具备借助思维链（Chain-of-Thought, CoT）进行推理的能力，当它们通过与内部人工“声音”对话来解释自身推理过程时，便为学习者提供了绝佳的共学契机，这是学习贝叶斯推理的一种极具吸引力的优质途径。此外，由于不同LLM有时会得出截然相反的结论，CoT还能通过对推理过程的细致比对，为深度学习创造机会。但我们发现，与人类不同的是，LLM不会主动采用自然频率、整体对象、具身启发式等具备生态效度的策略进行解释。这一情况令人遗憾，因为这些策略能帮助人类规避关键错误，且在贝叶斯推理教学中已被证实具有重要价值。为克服这些偏差并辅助理解与学习，我们加入了引导LLM使用上述策略的提示词。结果显示，采用CoT的LLM会纳入这些策略，但表现并不稳定。它们对符号推理存在持续性偏差，并且对生态效度策略存在回避或“恐惧症”。
英文摘要：
Learning to reason and carefully explain arguments is central to students' cognitive, mathematical, and computational thinking development. This is particularly challenging in problems under uncertainty and in Bayesian reasoning. With the new generation of large language models (LLMs) capable of reasoning using Chain-of-Thought (CoT), there is an excellent opportunity to learn with them as they explain their reasoning through a dialogue with their artificial internal voice. It is an engaging and excellent opportunity to learn Bayesian reasoning. Furthermore, given that different LLMs sometimes arrive at opposite solutions, CoT generates opportunities for deep learning by detailed comparisons of reasonings. However, unlike humans, we found that they do not autonomously explain using ecologically valid strategies like natural frequencies, whole objects, and embodied heuristics. This is unfortunate, as these strategies help humans avoid critical mistakes and have proven pedagogical value in Bayesian reasoning. In order to overcome these biases and aid understanding and learning, we included prompts that induce LLMs to use these strategies. We found that LLMs with CoT incorporate them but not consistently. They show persistent biases towards symbolic reasoning and avoidance or phobia of ecologically valid strategies.
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。
