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 2310.00378cs.CLcs.AIcs.CY
收藏

ValueDCG:衡量语言模型的综合人类价值观理解能力

ValueDCG: Measuring Comprehensive Human Value Understanding Ability of Language Models

Institute for Artificial Intelligence, Peking University(北京大学人工智能研究院);State Key Laboratory of General Artificial Intelligence, BIGAI(北京通用人工智能研究院通用人工智能国家重点实验室);School of Artificial Intelligence, Beijing University of Posts and Telecommunications(北京邮电大学人工智能学院)

Zhaowei Zhang, Fengshuo Bai, Jun Gao, Yaodong Yang

AI总结:

本文提出 ValueDCG 指标,从“知道是什么”和“知道为什么”两方面评估 LLM 的人类价值观理解能力,发现参数量增长并未同步提升两类能力,反而可能带来表面解释风险。

摘要:

个人价值观是人类决策背后的关键因素。鉴于大型语言模型(LLM)已被证明会显著影响人类决策,确保它们准确理解人类价值观对于保障其安全性至关重要。然而,由于价值观具有复杂且适应性强的特性,评估模型对这些价值观的掌握情况并不容易。我们认为,LLM真正理解价值观需要同时兼顾“知道是什么”和“知道为什么”。为此,我们提出了一项综合评估指标 ValueDCG(Value Discriminator-Critique Gap,价值观判别—批判差距),并通过工程实现来定量评估这两个方面。我们评估了四个具有代表性的 LLM,并提供了有力证据:LLM在“知道是什么”和“知道为什么”能力上的增长率并不随参数量增加而同步提升,这导致随着参数量增大,模型理解人类价值观的能力反而下降。这可能进一步表明,LLM或许会依据所提供的上下文编造看似合理的解释,却并未真正理解其内在价值,从而揭示出潜在风险。

英文摘要:

Personal values are a crucial factor behind human decision-making. Considering that Large Language Models (LLMs) have been shown to impact human decisions significantly, it is essential to make sure they accurately understand human values to ensure their safety. However, evaluating their grasp of these values is complex due to the value's intricate and adaptable nature. We argue that truly understanding values in LLMs requires considering both "know what" and "know why". To this end, we present a comprehensive evaluation metric, ValueDCG (Value Discriminator-Critique Gap), to quantitatively assess the two aspects with an engineering implementation. We assess four representative LLMs and provide compelling evidence that the growth rates of LLM's "know what" and "know why" capabilities do not align with increases in parameter numbers, resulting in a decline in the models' capacity to understand human values as larger amounts of parameters. This may further suggest that LLMs might craft plausible explanations based on the provided context without truly understanding their inherent value, indicating potential risks.

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。

↑