Position: Stop Anthropomorphizing Intermediate Tokens as Reasoning/Thinking Trac

Researchers at Arizona State University argue that calling intermediate tokens generated by language models 'reasoning traces' or 'thinking traces' anthropomorphizes the models and misleads users about their capabilities. In a position paper, they present evidence that such language confuses the nature of AI systems and leads to questionable research practices. The authors call on the AI community to avoid anthropomorphizing these tokens.

Computer Science Artificial Intelligence Submitted on 14 Apr 2025 v1 https://arxiv.org/abs/2504.09762v1 , last revised 9 Jun 2026 this version, v4 Title:Position: Stop Anthropomorphizing Intermediate Tokens as Reasoning/Thinking Traces View PDF /pdf/2504.09762 HTML experimental https://arxiv.org/html/2504.09762v4 Abstract:Intermediate token generation ITG , where a model produces output before the solution, has become a standard method to improve the performance of language models on reasoning tasks. These intermediate tokens have been called \say{reasoning traces} or even \say{thinking traces} -- implicitly anthropomorphizing the traces, and implying that these traces resemble steps a human might take when solving a challenging problem, and as such can provide an interpretable window into the operation of the model's thinking process to the end user. In this position paper, we present evidence that this anthropomorphization isn't a harmless metaphor, and instead is quite dangerous -- it confuses the nature of these models and how to use them effectively, and leads to questionable research. We call on the community to avoid such anthropomorphization of intermediate tokens. Submission history From: Subbarao Kambhampati view email /show-email/984a0c5e/2504.09762 Mon, 14 Apr 2025 00:03:34 UTC 381 KB v1 /abs/2504.09762v1 Tue, 27 May 2025 16:35:47 UTC 536 KB v2 /abs/2504.09762v2 Fri, 6 Mar 2026 14:36:07 UTC 1,321 KB v3 /abs/2504.09762v3 v4 Tue, 9 Jun 2026 20:08:39 UTC 4,278 KB References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .