通过激活空间干预，让大模型自动推理而不依赖提示词。
Uncovering Latent Chain of Thought Vectors in Language Models
- 用模型激活向量构造引导信号，在推理时注入模型。
- 在GSM8k等多基准测试中表现优于传统提示法。
- 为调试模型行为提供新思路，适合研究者参考。
本文研究语言模型（LM）激活空间中的定向扰动能否编码复杂推理模式。通过在推理阶段注入从模型激活中提取的引导向量，我们检验这些向量是否能在无需自然语言提示的情况下诱导出链式思维（CoT）推理。实验在Llama3 8B Instruct和Mistral 7B v0.2 Instruct上进行，结果表明，激活空间干预在GSM8k、MMLU、AGI Eval和ARC AI2等多个推理基准上达到甚至超过传统CoT提示的效果。这表明神经网络激活中确实蕴含推理模式，为激活空间操控提供了新的应用方向。
原文摘要 · Abstract (English)
In this work, we examine how targeted perturbations in the activation space of Language Models (LMs) can encode complex reasoning patterns. We inject steering vectors, derived from LM activations, into LMs during inference time and study whether these vectors can induce Chain-of-Thought (CoT) reasoning in LMs without the need for natural language prompting. We demonstrate this approach on Llama3 8B Instruct and Mistral 7B v0.2 Instruct and show that activation-space interventions achieve competitive, if not superior, performance compared to traditional CoT prompting across multiple reasoning benchmarks, including GSM8k, MMLU, AGI Eval, and ARC AI2. These findings suggest that neural network activations can encode reasoning patterns, offering a new application of activation space manipulation as a tool for tuning model behavior.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。