arXiv 2504.071282026-01-15cs.CL

DeepSeek-R1 思维学：让我们思考LLM推理

DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning

Mila – Quebec AI Institute（Mila-魁北克人工智能研究所）；McGill University（麦吉尔大学）；University of Copenhagen（哥本哈根大学）

Sara Vera Marjanović, Arkil Patel, Vaibhav Adlakha, Milad Aghajohari, Parishad BehnamGhader, Mehar Bhatia, Aditi Khandelwal, Austin Kraft, Benno Krojer, Xing Ha… 展开作者

Sara Vera Marjanović, Arkil Patel, Vaibhav Adlakha, Milad Aghajohari, Parishad BehnamGhader, Mehar Bhatia, Aditi Khandelwal, Austin Kraft, Benno Krojer, Xing Han Lù, Nicholas Meade, Dongchan Shin, Amirhossein Kazemnejad, Gaurav Kamath, Marius Mosbach, Karolina Stańczak, Siva Reddy

备注：: 135 pages, Published to TMLR

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

DeepSeek-R1通过多步骤推理链实现复杂问题解决，研究揭示其推理性能的'甜点'及潜在安全漏洞。

摘要：

大型推理模型如DeepSeek-R1标志着LLM处理复杂问题方式的根本转变。与直接为给定输入生成答案不同，DeepSeek-R1会创建详细的多步骤推理链，似乎在提供答案前会

英文摘要：

Large Reasoning Models like DeepSeek-R1 mark a fundamental shift in how LLMs approach complex problems. Instead of directly producing an answer for a given input, DeepSeek-R1 creates detailed multi-step reasoning chains, seemingly "thinking" about a problem before providing an answer. This reasoning process is publicly available to the user, creating endless opportunities for studying the reasoning behaviour of the model and opening up the field of Thoughtology. Starting from a taxonomy of DeepSeek-R1's basic building blocks of reasoning, our analyses on DeepSeek-R1 investigate the impact and controllability of thought length, management of long or confusing contexts, cultural and safety concerns, and the status of DeepSeek-R1 vis-à-vis cognitive phenomena, such as human-like language processing and world modelling. Our findings paint a nuanced picture. Notably, we show DeepSeek-R1 has a 'sweet spot' of reasoning, where extra inference time can impair model performance. Furthermore, we find a tendency for DeepSeek-R1 to persistently ruminate on previously explored problem formulations, obstructing further exploration. We also note strong safety vulnerabilities of DeepSeek-R1 compared to its non-reasoning counterpart, which can also compromise safety-aligned 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。