Large language models (LLMs) can solve problems step-by-step. While this
chain-of-thought (CoT) reasoning boosts LLMs' performance, it is unclear if
LLMs \textit{know} when to use CoT and whether those CoT are always necessary
to answer the question. This paper shows that LLMs tend to generate redundant
calculations and reasoning on a manually constructed math QA dataset,
GSM8K-Zero. GSM8K-Zero is constructed such that the questions can be answered
without any calculations, but LLMs, including Llama-2 models and Claude-2, tend
to generate lengthy and unnecessary calculations to answer the questions. We
also conduct experiments to explain why LLMs generate redundant calculations
and reasonings. GSM8K-Zero is publicly available at
this https URL and
this https URL

LLMs tend to generate lengthy and unnecessary calculations on the math QA dataset GSM8K-Zero, even though the questions can be answered without any calculations.

大型语言模型的过度推理和冗余计算

Over-Reasoning and Redundant Calculation of Large Language Models

Most of the calculations in standard sphere decoders are redundant, in the
sense that they either calculate quantities that are never used or calculate
some quantities more than once. A new method, which is applicable to lattices
as well as finite constellations, is proposed to avoid these redundant
calculations while still returning the same result. Pseudocode is given to
facilitate immediate implementation. Simulations show that the speed gain with
the proposed method increases linearly with the lattice dimension. At dimension
60, the new algorithms avoid about 75% of all floating-point operations.

本研究提出一种新的方法，避免冗余计算，应用于球形解码器和格点，演示该算法在 60 维时可避免约 75% 的浮点运算。