Is In-Context Learning a Type of Gradient-Based Learning? Evidence from the Inverse Frequency Effect in Structural Priming
Abstract: Large language models (LLMs) have shown the emerging capability of in-context learning (ICL). One line of research has explained ICL as functionally performing gradient descent. In this paper, we introduce a new way of diagnosing whether ICL is functionally equivalent to gradient-based learning. Our approach is based on the inverse frequency effect (IFE) -- a phenomenon in which an error-driven learner is expected to show larger updates when trained on infrequent examples than frequent ones. The IFE has previously been studied in psycholinguistics because humans show this effect in the context of structural priming (the tendency for people to produce sentence structures they have encountered recently); the IFE has been used as evidence that human structural priming must involve error-driven learning mechanisms. In our experiments, we simulated structural priming within ICL and found that LLMs display the IFE, with the effect being stronger in larger models. We conclude that ICL is indeed a type of gradient-based learning, supporting the hypothesis that a gradient component is implicitly computed in the forward pass during ICL. Our results suggest that both humans and LLMs make use of gradient-based, error-driven processing mechanisms.
Paper Type: Long
Research Area: Linguistic theories, Cognitive Modeling and Psycholinguistics
Research Area Keywords: In-Context Learning, Large Language Models, Structural Priming, Inverse Frequency Effect, Computational Psycholinguistics
Contribution Types: Model analysis & interpretability
Languages Studied: English
Submission Number: 4132
Loading