The Problem With Using LLMs for Simple Decisions
中文摘要
大语言模型基于逐词预测机制，这使其在处理简单决策时存在局限性。
English Summary
LLMs are built on next-token prediction, making them suboptimal for simple decision-making.
原文节选
Most modern AI systems are built around next-token prediction. Continue reading on Medium »