语言模型如何生成俚语：人类与机器生成俚语用法的系统性比较
How do Language Models Generate Slang: A Systematic Comparison between Human and Machine-Generated Slang Usages
AI总结：
本文系统比较人类与GPT-4o、Llama-3生成的俚语用法，从偏差特征、创造性和信息量三方面评估，发现LLM虽掌握俚语创造性知识但与人类认知存在显著偏差，尚不足以支撑语言分析等外推任务。
摘要：
俚语是一种常用的非正式语言类型，对自然语言处理（NLP）系统构成了严峻挑战。然而，大语言模型（LLMs）的最新进展使这一问题变得更易处理。尽管LLM智能体正被越来越广泛地应用于俚语检测和俚语解释等中间任务，但其泛化能力和可靠性在很大程度上取决于这些模型是否已经掌握了与人类实际使用的俚语用法高度一致的结构性知识。为了回答这一问题，我们提出了一项人类与机器生成俚语用法的系统性比较。我们的评估框架聚焦于三个核心方面：1）反映机器如何感知俚语的系统性偏差的用法特征；2）俚语用法所体现的词汇新造和词语复用方面的创造性；3）俚语用法在作为模型蒸馏的金标准示例时所具有的信息量。通过比较来自在线俚语词典（OSD）的人类实际俚语用法与GPT-4o和Llama-3生成的俚语，我们发现LLM在感知俚语方面存在显著偏差。我们的结果表明，尽管LLM已经掌握了大量关于俚语创造性方面的知识，但这些知识与人类的契合程度尚不足以支持LLM执行语言分析等外推性任务。
英文摘要：
Slang is a commonly used type of informal language that poses a daunting challenge to NLP systems. Recent advances in large language models (LLMs), however, have made the problem more approachable. While LLM agents are becoming more widely applied to intermediary tasks such as slang detection and slang interpretation, their generalizability and reliability are heavily dependent on whether these models have captured structural knowledge about slang that align well with human attested slang usages. To answer this question, we contribute a systematic comparison between human and machine-generated slang usages. Our evaluative framework focuses on three core aspects: 1) Characteristics of the usages that reflect systematic biases in how machines perceive slang, 2) Creativity reflected by both lexical coinages and word reuses employed by the slang usages, and 3) Informativeness of the slang usages when used as gold-standard examples for model distillation. By comparing human-attested slang usages from the Online Slang Dictionary (OSD) and slang generated by GPT-4o and Llama-3, we find significant biases in how LLMs perceive slang. Our results suggest that while LLMs have captured significant knowledge about the creative aspects of slang, such knowledge does not align with humans sufficiently to enable LLMs for extrapolative tasks such as linguistic analyses.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
