我们研究大型语言模型(LLMs)是否在复杂提示下潜在地执行多跳推理,例如“The mother of the singer of 'Superstition' is”。我们寻找潜在推理路径的证据,其中LLM(1)潜在地将“the singer of 'Superstition'”识别为桥梁实体Stevie Wonder,以及(2)利用其关于Stevie Wonder母亲的知识来完成提示。我们分别分析这两个跳,并将它们的共同出现视为潜在多跳推理的指示。对于第一跳,我们测试将提示更改为间接提及桥梁实体而非任何其他实体,是否会增加LLM对桥梁实体的内部回忆。对于第二跳,我们测试增加这种回忆是否会导致LLM更好地利用其关于桥梁实体的知识。我们发现,对于某些关系类型的提示,存在潜在多跳推理的有力证据,该推理路径在超过80%的提示中被使用。然而,这种利用高度依赖上下文,在不同类型的提示之间存在差异。此外,平均而言,第二跳和完整多跳遍历的证据相当温和,仅第一跳的证据较为显著。此外,我们发现随着模型规模增加,第一跳推理存在明显的扩展趋势,但第二跳则没有。我们的实验发现为LLMs的未来发展和应用提出了潜在的挑战和机遇。
英文摘要:
We study whether Large Language Models (LLMs) latently perform multi-hop reasoning with complex prompts such as "The mother of the singer of 'Superstition' is". We look for evidence of a latent reasoning pathway where an LLM (1) latently identifies "the singer of 'Superstition'" as Stevie Wonder, the bridge entity, and (2) uses its knowledge of Stevie Wonder's mother to complete the prompt. We analyze these two hops individually and consider their co-occurrence as indicative of latent multi-hop reasoning. For the first hop, we test if changing the prompt to indirectly mention the bridge entity instead of any other entity increases the LLM's internal recall of the bridge entity. For the second hop, we test if increasing this recall causes the LLM to better utilize what it knows about the bridge entity. We find strong evidence of latent multi-hop reasoning for the prompts of certain relation types, with the reasoning pathway used in more than 80% of the prompts. However, the utilization is highly contextual, varying across different types of prompts. Also, on average, the evidence for the second hop and the full multi-hop traversal is rather moderate and only substantial for the first hop. Moreover, we find a clear scaling trend with increasing model size for the first hop of reasoning but not for the second hop. Our experimental findings suggest potential challenges and opportunities for future development and applications of LLMs.
National Institute of Informatics(情报学研究所);Nantes Université(南特大学);The Graduate University for Advanced Studies(高级研究大学院大学);The University of Tokyo(东京大学);EPFL(洛桑联邦理工学院)
Unleashing Multi-Hop Reasoning Potential in Large Language Models through Repetition of Misordered Context
Department of Electrical and Computer Engineering, Seoul National University(电子与计算机工程系,首尔国立大学);AIIS, ASRI, INMC, ISRC, and IPAI, Seoul National University(人工智能研究所、人工智能研究机构、智能网络中心、智能系统研究所以及人工智能与公共政策研究所,首尔国立大学)
BELLE: A Bi-Level Multi-Agent Reasoning Framework for Multi-Hop Question Answering
School of Computer Science and Information Engineering, Hefei University of Technology(合肥工业大学计算机科学与信息工程学院);Shanghai University of Electric Power(上海电力大学);Alibaba Group(阿里巴巴集团);East China Normal University(华东师范大学)
Robust Knowledge Editing via Explicit Reasoning Chains for Distractor-Resilient Multi-Hop QA
Shanghai Jiao Tong University(上海交通大学);The University of Sydney(悉尼大学);Shenzhen Campus of Sun Yat-sen University(中山大学深圳校区);Nanyang Technological University(南洋理工大学)
Seemingly Plausible Distractors in Multi-Hop Reasoning: Are Large Language Models Attentive Readers?
ASUS Intelligent Cloud Services (AICS)(华硕智能云服务(AICS));National University of Singapore(新加坡国立大学);University of Manchester(曼彻斯特大学);Imperial Global Singapore(帝国理工学院新加坡分部)
Scaling Reasoning Hop Exposes Weaknesses: Demystifying and Improving Hop Generalization in Large Language Models
University of Science and Technology of China(中国科学技术大学);City University of Hong Kong(香港城市大学);Zhejiang University(浙江大学);City University of Hong Kong, Shenzhen Research Institute(香港城市大学深圳研究院)
Layer-Order Inversion: Rethinking Latent Multi-Hop Reasoning in Large Language Models
State Key Laboratory of Cognitive Intelligence(认知智能国家重点实验室);University of Science and Technology of China(中国科学技术大学);The Hong Kong University of Science and Technology(香港科技大学)
Understanding Reasoning in Chain-of-Thought from the Hopfieldian View
King Abdullah University of Science and Technology(阿卜杜拉国王科技大学);Soochow University(苏州大学);University of Southampton(南安普顿大学);Arizona State University(亚利桑那州立大学)
Effect of Document Packing on the Latent Multi-Hop Reasoning Capabilities of Large Language Models
Chandar Research Lab(Chandar研究实验室);Mila – Quebec AI Institute(魁北克人工智能研究所);Université de Montréal(蒙特利尔大学);Microsoft Research(微软研究院);Polytechnique Montréal(蒙特利尔理工学院);Canada CIFAR AI Chair(加拿大CIFAR人工智能主席)
A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning
KAUST(阿卜杜拉国王科技大学);Soochow University(苏州大学);University of Southampton(南安普顿大学);Tongji University(同济大学);Universität Hamburg(汉堡大学);Huazhong University of Science and Technology(华中科技大学)
Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Steps
National Institute of Advanced Industrial Science and Technology(产业技术综合研究所);National Institute of Informatics(国立信息学研究所);The Graduate University for Advanced Studies(综合研究大学院大学)