Stop Anthropomorphisizing Intermediate Tokens: Qwen3.8 doesn't "overthink"

Intermediate tokens, called "thinking" or "reasoning" actually are nothing like it. Humans do step-by-step reasoning leading to the conclusion. LLMs use intermediate traces to augment their prompt. This explains why sometimes the answer is very good but the "reasoning" is verbose. Flooding you

中文全文翻译正在补齐，可先查看下方英文原文。

Intermediate tokens, called "thinking" or "reasoning" actually are nothing like it. Humans do step-by-step reasoning leading to the conclusion. LLMs use intermediate traces to augment their prompt. This explains why sometimes the answer is very good but the "reasoning" is verbose. Flooding you