arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2403.18802cs.CLcs.AIcs.LG
收藏

大型语言模型中的长文本事实性

Long-form factuality in large language models

Google DeepMind(谷歌DeepMind);Stanford University(斯坦福大学);University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)

Jerry Wei, Chengrun Yang, Xinying Song, Yifeng Lu, Nathan Hu, Jie Huang, Dustin Tran, Daiyi Peng, Ruibo Liu, Da Huang, Cosmo Du, Quoc V. Le

AI总结:

本文提出LongFact基准和SAFE评估方法,利用LLM智能体自动评估长文本事实性,并扩展F1指标,实验表明SAFE优于人类标注者且成本更低,大型模型事实性更好。

摘要:

大型语言模型(LLM)在回答开放式主题的事实性提示时,常常生成包含事实错误的内容。为了在开放领域对模型的长文本事实性进行基准测试,我们首先使用GPT-4生成LongFact,这是一个包含数千个问题、涵盖38个主题的提示集。然后,我们提出LLM智能体可以通过一种我们称之为搜索增强事实性评估器(SAFE)的方法用作长文本事实性的自动评估器。SAFE利用LLM将长文本响应分解为一组独立的事实,并通过多步推理过程评估每个事实的准确性,该过程包括向Google搜索发送查询并确定事实是否得到搜索结果的支持。此外,我们提出将F1分数扩展为长文本事实性的聚合指标。为此,我们平衡响应中受支持事实的百分比(精确率)与相对于代表用户偏好响应长度的超参数所提供事实的百分比(召回率)。在实验上,我们证明LLM智能体可以胜过众包人类标注者——在约1.6万个独立事实的集合上,SAFE与人类标注者的一致性达到72%,在随机选择的100个分歧案例中,SAFE在76%的情况下胜出。同时,SAFE的成本比人类标注者低20倍以上。我们还在四个模型家族(Gemini、GPT、Claude和PaLM-2)的LongFact上对十三个语言模型进行了基准测试,发现较大的语言模型通常能实现更好的长文本事实性。LongFact、SAFE及所有实验代码可在https://github.com/google-deepmind/long-form-factuality获取。

英文摘要:

Large language models (LLMs) often generate content that contains factual errors when responding to fact-seeking prompts on open-ended topics. To benchmark a model's long-form factuality in open domains, we first use GPT-4 to generate LongFact, a prompt set comprising thousands of questions spanning 38 topics. We then propose that LLM agents can be used as automated evaluators for long-form factuality through a method which we call Search-Augmented Factuality Evaluator (SAFE). SAFE utilizes an LLM to break down a long-form response into a set of individual facts and to evaluate the accuracy of each fact using a multi-step reasoning process comprising sending search queries to Google Search and determining whether a fact is supported by the search results. Furthermore, we propose extending F1 score as an aggregated metric for long-form factuality. To do so, we balance the percentage of supported facts in a response (precision) with the percentage of provided facts relative to a hyperparameter representing a user's preferred response length (recall). Empirically, we demonstrate that LLM agents can outperform crowdsourced human annotators - on a set of ~16k individual facts, SAFE agrees with crowdsourced human annotators 72% of the time, and on a random subset of 100 disagreement cases, SAFE wins 76% of the time. At the same time, SAFE is more than 20 times cheaper than human annotators. We also benchmark thirteen language models on LongFact across four model families (Gemini, GPT, Claude, and PaLM-2), finding that larger language models generally achieve better long-form factuality. LongFact, SAFE, and all experimental code are available at https://github.com/google-deepmind/long-form-factuality.

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。

↑