研究提出 Transformer 错误率的两参数模型

View PDF HTML (experimental)

Abstract:We study the error rate of LLMs on tasks like arithmetic that require a deterministic output, and repetitive processing of tokens drawn from a small set of alternatives. We argue that incorrect predictions arise when small errors in the attention mechanism accumulate to cross a threshold, and use this insight to derive a quantitative two-parameter relationship between the accuracy and the complexity of the task. The two parameters vary with the prompt and the model; they can be interpreted in terms of an elementary noise rate, and the number of plausible erroneous tokens that can be predicted. Our analysis is inspired by an ``effective field theory'' perspective: the LLM's many raw parameters can be reorganized into just two parameters that govern the error rate. We perform extensive empirical tests, using Gemini 2.5 Flash, Gemini 2.5 Pro and DeepSeek R1, and find excellent agreement between the predicted and observed accuracy for a variety of tasks, although we also identify deviations in some cases. Our model provides an alternative to suggestions that errors made by LLMs on long repetitive tasks indicate the ``collapse of reasoning'', or an inability to express ``compositional'' functions. Finally, we show how to construct prompts to reduce the error rate.

Comments:	(v1) 8+17pages (v2) Appendix A.4 expanded; consistent with published version
Subjects:	Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); High Energy Physics - Theory (hep-th)
ACM classes:	I.2.7; I.2.6
Cite as:	arXiv:2601.14175 [cs.LG]
	(or arXiv:2601.14175v2 [cs.LG] for this version)
	https://doi.org/10.48550/arXiv.2601.14175 arXiv-issued DOI via DataCite
Journal reference:	Proceedings of the 43rd International Conference on Machine Learning, Seoul, South Korea. PMLR 306, 2026

Submission history

From: Suvrat Raju [view email]
[v1] Tue, 20 Jan 2026 17:27:03 UTC (146 KB)
[v2] Thu, 12 Nov 2026 13:22:46 UTC (203 KB)