On the Branching Bias of Syntax Extracted from Pre-trained Language Models

Abstract: Many efforts have been devoted to extracting constituency trees from pre-trained LLMs, often proceeding in two stages: feature definition and parsing. However, this kind of methods may suffer from the branching bias issue, which will inflate the performances on languages with the same branch it biases to. In this work, we propose quantitatively measuring the branching bias by comparing the performance gap on a language and its reversed language, which is agnostic to both LLMs and extracting methods. Furthermore, we analyze the impacts of three factors on the branching bias, namely parsing algorithms, feature definitions, and LLMs. Experiments show that several existing works exhibit branching biases, and some implementations of these three factors can introduce the branching bias.