Over-Reasoning and Redundant Calculation of Large Language Models
This identifies a specific inefficiency in LLM reasoning for math QA, which is incremental but highlights a limitation in current models' ability to assess when reasoning is needed.
The paper tackles the problem of large language models (LLMs) generating unnecessary step-by-step reasoning, even when questions can be answered directly without calculations, as demonstrated on a manually constructed dataset GSM8K-Zero, where models like Llama-2 and Claude-2 produced redundant calculations.
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 https://github.com/d223302/Over-Reasoning-of-LLMs and https://huggingface.co/datasets/dcml0714/GSM8K-Zero.