通过思维链提示评估大语言模型中的性别偏见
Evaluating Gender Bias in Large Language Models via Chain-of-Thought Prompting
AI总结：
本研究针对不可扩展任务中CoT提示对大语言模型性别偏见的影响展开探究，构建了兼具算术与符号推理特征的词汇计数基准，发现CoT提示可减轻大语言模型的无意识社会偏见，促进公平预测。
摘要：
存在两类任务：一类是可扩展任务，如阅读理解和事实验证，模型性能随模型规模提升而提高；另一类是不可扩展任务，如算术推理和符号推理，模型性能不一定随模型规模提升而提高。配备Chain-of-Thought（CoT，思维链）提示的大语言模型（LLM）即使在不可扩展任务上也能做出准确的增量预测。遗憾的是，尽管大语言模型具备出色的推理能力，它们往往会内化并复现具有歧视性的社会偏见。对于不可扩展任务中的隐含信息，CoT能否给出歧视性或平等主义的合理化解释仍是一个悬而未决的问题。\n在本研究中，我们考察了大语言模型的逐步预测对不可扩展任务中性别偏见的影响。为此，我们构建了一个面向不可扩展任务的基准：向大语言模型提供包含女性词汇、男性词汇以及带性别属性的职业词汇的词表，要求其统计女性词汇和男性词汇的数量。在我们的CoT提示中，我们要求大语言模型在做出最终预测前，明确指出词表中的每个单词是女性词汇还是男性词汇。该基准涉及词汇计数和词义处理，兼具算术推理和符号推理的特征。英文语境下的实验结果表明，在没有逐步预测的情况下，尽管任务简单如词汇计数，大多数大语言模型仍会做出带有社会偏见的预测。有趣的是，CoT提示减轻了大语言模型中的这种无意识社会偏见，促进了公平预测。
英文摘要：
There exist both scalable tasks, like reading comprehension and fact-checking, where model performance improves with model size, and unscalable tasks, like arithmetic reasoning and symbolic reasoning, where model performance does not necessarily improve with model size. Large language models (LLMs) equipped with Chain-of-Thought (CoT) prompting are able to make accurate incremental predictions even on unscalable tasks. Unfortunately, despite their exceptional reasoning abilities, LLMs tend to internalize and reproduce discriminatory societal biases. Whether CoT can provide discriminatory or egalitarian rationalizations for the implicit information in unscalable tasks remains an open question. In this study, we examine the impact of LLMs' step-by-step predictions on gender bias in unscalable tasks. For this purpose, we construct a benchmark for an unscalable task where the LLM is given a list of words comprising feminine, masculine, and gendered occupational words, and is required to count the number of feminine and masculine words. In our CoT prompts, we require the LLM to explicitly indicate whether each word in the word list is a feminine or masculine before making the final predictions. With counting and handling the meaning of words, this benchmark has characteristics of both arithmetic reasoning and symbolic reasoning. Experimental results in English show that without step-by-step prediction, most LLMs make socially biased predictions, despite the task being as simple as counting words. Interestingly, CoT prompting reduces this unconscious social bias in LLMs and encourages fair predictions.
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。
