Do Language Models Exhibit Human-like Structural Priming Effects?

TL;DR

This paper investigates whether language models exhibit human-like structural priming effects, analyzing how sentence and token-level factors influence predictions and comparing these effects to human data.

Contribution

It demonstrates that language models show priming effects driven by inverse frequency and lexical dependence, aligning with human priming mechanisms.

Findings

Priming effects are influenced by inverse frequency of elements.

Lexical dependence between prime and target affects priming.

Language models exhibit effects similar to human structural priming.

Abstract

We explore which linguistic factors -- at the sentence and token level -- play an important role in influencing language model predictions, and investigate whether these are reflective of results found in humans and human corpora (Gries and Kootstra, 2017). We make use of the structural priming paradigm, where recent exposure to a structure facilitates processing of the same structure. We don't only investigate whether, but also where priming effects occur, and what factors predict them. We show that these effects can be explained via the inverse frequency effect, known in human priming, where rarer elements within a prime increase priming effects, as well as lexical dependence between prime and target. Our results provide an important piece in the puzzle of understanding how properties within their context affect structural prediction in language models.