aifollow.news
返回 Rohan Paul
Rohan Paul· @rohanpaul_ai · X· · 原发布时间 AI 评分54

CPU-ARC 研究提出将重复推理转为 CPU 代码复用

AI 导读

一篇关于 CPU-ARC 的论文提出:让 Claude Opus 4.5 将推理过程转成可接收不同输入的 Python 函数,后续在 CPU 上执行,仅在提取数值等小型文本任务中调用微调后的 1B 模型。所引介绍称,该方法在 8 项数学和科学基准中的 7 项达到或超过源模型准确率;这些结果尚不能说明它适用于所有重复任务。

原帖引用 · 原文

@rohanpaul_ai

– https://arxiv.org/abs/2611.09891

Title: "CPU-ARC: Agentic Reasoning Caching for on-CPU Reasoning Reuse"

回复Rohan Paul@rohanpaul_ai
New NVIDIA paper finds that once an LLM has reasoned through a problem, the solution can run as CPU code for rephrased and re-numbered versions of it. that answers similar questions 16x to 291x cheaper. LLMs re-solve every problem from scratch, even when only wording or numbers change. Agents sending the same instructions with new data pay for that repeated reasoning each time. Their method has Claude Opus 4.5 turn a reasoning trace into small Python functions, with changeable inputs as parameters. At run time they execute on a CPU, calling a fine-tuned 1B model only for small text jobs like extracting values. The cache matched or beat the source model's accuracy on 7 of 8 math and science benchmarks, scoring 93.3% versus 83.3% on AIME '25. It handled at least 90% of rephrased or re-numbered variants. If your agent runs the same prompt on new inputs all day, cache the reasoning as code and use the LLM only on misses.
在 X 查看这条帖子 ↗

来源:Rohan Paul · x.com

论文
发现内容有误?提交纠错