通过表征工程激发大模型通用长链推理能力
Unlocking General Long Chain-of-Thought Reasoning Capabilities of Large Language Models via Representation Engineering
- 从表征角度分析，发现大模型具备通用长链推理能力
- 仅需少量示例微调，即可在跨领域任务中有效迁移
- 适合希望提升模型推理泛化能力的研究者
近期长链思维（long CoT）进展显著提升了大语言模型（LLMs）的推理能力。现有研究发现，仅需少量示例微调即可高效激发长链推理能力，并可轻松迁移到其他任务。这促使我们探究长链推理是否为大模型的通用能力。本文从表征视角进行实证分析，发现大模型确实编码了长链推理的通用能力，且与普通链式思维有明显区别。此外，领域特定表征对长链推理的有效迁移至关重要。基于此，我们提出 GLoRE，一种新型表征工程方法，以释放大模型的通用长链推理潜力。大量实验表明，GLoRE在同域和跨域场景下均具有高效性与有效性。
原文摘要 · Abstract (English)
Recent advancements in long chain-of-thoughts(long CoTs) have significantly improved the reasoning capabilities of large language models(LLMs). Existing work finds that the capability of long CoT reasoning can be efficiently elicited by tuning on only a few examples and can easily transfer to other tasks. This motivates us to investigate whether long CoT reasoning is a general capability for LLMs. In this work, we conduct an empirical analysis for this question from the perspective of representation. We find that LLMs do encode long CoT reasoning as a general capability, with a clear distinction from vanilla CoTs. Furthermore, domain-specific representations are also required for the effective transfer of long CoT reasoning. Inspired by these findings, we propose GLoRE, a novel representation engineering method to unleash the general long CoT reasoning capabilities of LLMs. Extensive experiments demonstrate the effectiveness and efficiency of GLoRE in both in-domain and cross-domain scenarios.
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。