JI Z, LEE N, FRIESKE R, et al. Survey of Hallucination in Natural Language Generation[J/OL]. ACM Computing Surveys, 2023: 1-38.http://dx.doi.org/10.1145/3571730.DOI:10.1145/3571730. 1 intro 在自然语言处理问题中,研究者观察到NLG模型中的训练目标可能导致训练存在缺陷,即生成结果输出乏味、...
在自然语言处理领域，研究者观察到自然语言生成（NLG）模型的训练目标可能引发生成结果的缺陷，导致输出显得乏味、不连贯或陷入循环。同时，模型有时会生成无意义的文本或与提供的输入信息不符，这一现象被称为幻觉问题（hallucination）。本文从幻觉问题出发，探讨不同任务下的具体问题，如抽象总结、对话生成...
未来方向 有一篇最近的高引用论文《Survey ofHallucinationin Natural Language Generation》,是香港科技大学的成果。 讨论的正好就是大模型的幻觉问题,也就是Hallucination。 幻觉的定义 先说幻觉Hallucination的定义,作者采用了这个定义: the generated content that is nonsensical or unfaithful to the provided source ...
spss=adap_pc [7]Tonmoy S M, Zaman S M, Jain V, et al. A Comprehensive Survey of Hallucination Mitigation Techniques in Large Language Models[J]. arXiv preprint arXiv:2401.01313, 2024 [8]Ji Z, Lee N, Frieske R, et al. Survey of hallucination in natural language generation[J]. ACM ...
Natural Language Generation (NLG) has improved exponentially in recent years thanks to the development of sequence-to-sequence deep learning technologies such as Transformer-based language models. This advancement has led to more fluent and coherent NLG, leading to improved development in downstream task...
Generally, hallucinations in natural language generation tasks can be categorized into two primary types: intrinsic hallucination and extrinsic hallucination [126, 136, 174]. 幻觉的概念溯源于病理学和心理学领域,被定义为对现实中缺席的实体或事件的感知[ 202 ]。在NLP领域,幻觉通常被认为是一种产生的内容...
Survey of hallucination in natural language generation. ACM Comput. Surv., 55(12), mar 2023a. ISSN 0360-0300. doi: 10.1145/3571730. URL https://doi.org/10.1145/3571730. Ji et al. (2023b) Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye...
LANGUAGE modelsHALLUCINATIONSRESEARCH personnelDespite impressive advances in Natural Language Generation (NLG) and Large Language Models (LLMs), researchers are still unclear about important aspects of NLG evaluation. To substantiate this claim, I examine current classifications of hallucination and omission...
nlp machine-learning web-crawler llama datasets dataset-generation system-design hallucination rag large-language-models llm Updated Mar 18, 2024 HTML dmis-lab / OLAPH Star 36 Code Issues Pull requests OLAPH: Improving Factuality in Biomedical Long-form Question Answering question-answering hal...
The emergence of large language models (LLMs) has marked a significant breakthrough in natural language processing (NLP), fueling a paradigm shift in information acquisition. Nevertheless, LLMs are prone to hallucination, generating plausible yet nonfactual content. This phenomenon raises significant co...