生成语言模型中演绎推理的出现

May, 2023

Examining the Emergence of Deductive Reasoning in Generative Language Models

Peter Belcak, Luca A. Lanzendörfer, Roger Wattenhofer

TL;DR本研究初步探讨了生成 Transformer 模型从给定前提中演绎推理的能力，发现其性能与训练设置和参数规模有显著差异，而且推理链的长度一般不会影响性能，除了 OpenAI GPT-3 和 GPT-3.5 模型。研究考虑了大小从 1.17 亿到 1750 亿参数的多种变压器解码器模型。

Abstract

We conduct a preliminary inquiry into the ability of generative transformer models to deductively reason from premises provided. We observe notable differences in the performance of models coming from different <