ThoughtTerminator: Benchmarking, Calibrating, and Mitigating Overthinking in Reasoning Models

Xiao Pu, Michael Saxon, Wenyue Hua, William Yang Wang

Keywords: reasoning model, overthinking, decoding, tool use, evaluation, benchmark

TL;DR: We evaluate the relationship between problem difficulty and token cost in reasoning models, benchmark how efficiently different models allocate tokens, and introduce a simple training-free decoding method to reduce overthinking, Thought Terminator.

Abstract: Reasoning models have demonstrated impressive performance on difficult tasks that traditional language models struggle at. However, many are plagued with the problem of overthinking---generating large amounts of unnecessary tokens which don't improve accuracy on a question. We introduce approximate measures of problem-level difficulty and demonstrate that a clear relationship between problem difficulty and optimal token spend exists, and evaluate how well calibrated a variety of reasoning models are in terms of efficiently allocating the optimal token count. We find that in general, reasoning models are poorly calibrated, particularly on easy problems. To evaluate calibration on easy questions we introduce DUMB500, a dataset of extremely easy math, reasoning, code, and task problems, and jointly evaluate reasoning model on these simple examples and extremely difficult examples from existing frontier benchmarks on the same task domain. Finally, we introduce ThoughtTerminator, a training-free black box decoding technique that significantly improves reasoning model calibration.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the COLM Code of Ethics on https://colmweb.org/CoE.html

Author Guide: I certify that this submission complies with the submission instructions as described on https://colmweb.org/AuthorGuide.html

Submission Number: 536