Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times?

Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times? Byung-Doh Oh Department of Linguistics The Ohio State University, USA oh.531@osu.edu William Schuler Department of Linguistics The Ohio State University, USA schuler.77@osu.edu Abstract This work presents a linguistic analysis into why larger Transformer-based pre-trained lan- guage models with more parameters and lower perplexity nonetheless yield surprisal estimates that are less predictive of human reading times. First, regression analyses show a strictly monotonic, positive log-linear rela- tionship between perplexity and fit to reading times for the more recently released five GPT-Neo variants and eight OPT variants on two separate datasets, replicating earlier results limited to just GPT-2 (Oh et al., 2022). Subsequently, analysis of residual er- rors reveals a systematic deviation of the larger variants, such as underpredicting reading times of named entities and making compensatory overpredictions for reading times of func- tion words such as modals and conjunctions. These results suggest that the propensity of larger Transformer-based models to ‘memo- rize’ sequences during training makes their surprisal estimates diverge from humanlike expectations, which warrants caution in using pre-trained language models to study human language processing. 1 Introduction Expectation-based theories of sentence processing (Hale, 2001; Levy, 2008) postulate that process- ing difficulty is largely driven by how predictable upcoming linguistic material is given its context. In cognitive modeling, predictability operational- ized by information-theoretic surprisal (Shannon, 1948) has been shown to be a strong predictor of behavioral and neural measures of process- ing difficulty (Demberg and Keller, 2008; Smith and Levy, 2013; Hale et al., 2018; Shain et al., 2020), providing empirical support for this posi- tion. As language models (LMs) directly define a conditional probability distribution of a word given its context required for surprisal calculation, they have frequently been evaluated as surprisal-based cognitive models of sentence processing. Recently, it was observed that surprisal from larger variants of the pre-trained GPT-2 LM (Radford et al., 2019) that have more parame- ters and achieve lower perplexity is less predictive of self-paced reading times and eye-gaze dura- tions collected during naturalistic reading (Oh et al., 2022). As the different variants of the pre- trained GPT-2 model share the primary architec- ture and training data, this offers an especially strong counterexample to previous work that showed a negative relationship between LM per- plexity and predictive power of surprisal esti- mates (Goodkind and Bicknell, 2018; Hao et al., 2020; Wilcox et al., 2020). More broadly, this observation also contradicts the recent ‘larger is better’ trend of the NLP community, leaving open the question of why larger LMs perform worse. However, the Oh et al. (2022) results were part of a follow-up analysis in support of a separate claim about parser surprisal that only examined four model variants, so the results were not tested for statistical significance or extensively explored. The current work fills that gap by conducting a detailed linguistic analysis of the positive re- lationship between LM perplexity and predictive power of surprisal estimates. First, the robustness of …

Content truncated to 500 words