TRACES: Tagging Reasoning Steps for Adaptive Cost-Efficient Early-Stopping

arXiv:2604.21057 · cs.CL · Submitted 2026-04-22 · Read on arXiv

cs.CL

Submitted: 2026-04-22

Updated: 2026-08-27

Comments: Accepted at the Third Conference on Language Modeling (COLM 2026)

Code: https://github.com/ZubinGou/math-evaluation-harness

License: http://creativecommons.org/licenses/by/4.0/

The gist: The field of Language Reasoning Models (LRMs) has been very active over the past few years with advances in training and inference techniques enabling LRMs to reason longer, and more accurately.

Terminology

Abstract

The field of Language Reasoning Models (LRMs) has been very active over the past few years with advances in training and inference techniques enabling LRMs to reason longer, and more accurately. However, a growing body of studies show that LRMs are still inefficient, over-generating verification and reflection steps. Additionally, the high-level role of each reasoning step and how different step types contribute to the generation of correct answers, is largely underexplored. To address this challenge, we introduce TRACES (Tagging of the Reasoning steps enabling Adaptive Cost-Efficient early-Stopping), a lightweight framework that tags reasoning steps in real-time, and enable adaptive, cost-efficient early stopping of large-language-model inferences. By monitoring reasoning behaviors during inferences, we find that LRMs tend to shift their reasoning behavior after reaching a correct answer. We demonstrate that the monitoring of the specific type of steps can produce effective interpretable early stopping criteria. Evaluated on three mathematical reasoning benchmarks (MATH500, GSM8K, AIME) and two knowledge and reasoning benchmarks (MMLU and GPQA), TRACES achieve 20 to 50% token reduction while maintaining comparable accuracy to standard generation. On harder tasks (BeyondAIME, IMO AnswerBench), the accuracy-efficiency trade-off is steeper, requiring a more conservative early-stopping threshold. This work offers a novel way to study and control the generation behavior of LRMs.

Sources

Related papers