Do Language Models Exhibit Human-like Structural Priming Effects?
This research addresses the puzzle of how context influences structural prediction in language models, which is incremental as it applies known human paradigms to AI.
The study investigated whether language models show structural priming effects similar to humans, finding that these effects can be explained by inverse frequency and lexical dependence, with rarer elements increasing priming.
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.