arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.12751cs.CL
收藏

字符级中文 Backpack 语言模型

Character-level Chinese Backpack Language Models

Stanford University(斯坦福大学)

Hao Sun, John Hewitt

AI总结:

在字符分词的中文上训练并评估 Backpack 语言模型,发现其性能与 Transformer 相当,字符语义能组合形成词义,且可通过定位并干预特定字符语义来减少性别偏差。

摘要:

Backpack 是一种 Transformer 替代方案,通过将预测分解为 token 语义组件的加权和,已被证明能提升英语语言建模的可解释性。然而,Backpack 依赖 token 定义的语义,这对其在英语以外语言的潜力提出了疑问,因为在英语中子词分词能为词汇项提供合理的近似。在本研究中,我们在字符分词的中文上训练、评估、解释和控制 Backpack 语言模型,在中文里词通常由多个字符组成。我们发现,我们的(134M 参数)中文 Backpack 语言模型表现与(104M 参数)的 Transformer 相当,并学习到丰富的字符级语义,这些语义以对数相加的方式组合形成词义。在 SimLex 风格的词汇语义评估中,Backpack 字符语义的简单平均优于 Transformer 的输入嵌入。我们发现,复杂的多字符语义通常是通过在不同上下文中一致地使用相同的每字符语义权重来形成的。通过控制探索可解释性,我们展示了可以将 Backpack 中性别偏差的来源定位到特定的字符语义,并进行干预以减少偏差。

英文摘要:

The Backpack is a Transformer alternative shown to improve interpretability in English language modeling by decomposing predictions into a weighted sum of token sense components. However, Backpacks' reliance on token-defined meaning raises questions as to their potential for languages other than English, a language for which subword tokenization provides a reasonable approximation for lexical items. In this work, we train, evaluate, interpret, and control Backpack language models in character-tokenized Chinese, in which words are often composed of many characters. We find that our (134M parameter) Chinese Backpack language model performs comparably to a (104M parameter) Transformer, and learns rich character-level meanings that log-additively compose to form word meanings. In SimLex-style lexical semantic evaluations, simple averages of Backpack character senses outperform input embeddings from a Transformer. We find that complex multi-character meanings are often formed by using the same per-character sense weights consistently across context. Exploring interpretability-through control, we show that we can localize a source of gender bias in our Backpacks to specific character senses and intervene to reduce the bias.

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。

↑