Toward In-Context Teaching: Adapting Examples to Students' Misconceptions

TL;DR

This paper introduces a new framework and models for adaptive teaching that tailor examples to students' misconceptions, demonstrating improved effectiveness over non-adaptive methods in both simulated and human studies.

Contribution

The paper presents AdapT, a comprehensive evaluation platform, and AToM, a probabilistic model for adaptive teaching, advancing the ability of models to personalize instruction.

Findings

AToM outperforms existing models in simulated learning tasks.

Both AToM and LLMs outperform random example selection in human experiments.

Adaptive models show promise but face challenges in real-world teaching scenarios.

Abstract

When a teacher provides examples for a student to study, these examples must be informative, enabling a student to progress from their current state toward a target concept or skill. Good teachers must therefore simultaneously infer what students already know and adapt their teaching to students' changing state of knowledge. There is increasing interest in using computational models, particularly large language models, as pedagogical tools. As students, language models in particular have shown a remarkable ability to adapt to new tasks given small numbers of examples. But how effectively can these models adapt as teachers to students of different types? To study this question, we introduce a suite of models and evaluation methods we call AdapT. AdapT has two components: (1) a collection of simulated Bayesian student models that can be used for evaluation of automated teaching methods; (2) a platform for evaluation with human students, to characterize the real-world effectiveness of these methods. We additionally introduce (3) AToM, a new probabilistic model for adaptive teaching that jointly infers students' past beliefs and optimizes for the correctness of future beliefs. In evaluations of simulated students across three learning domains (fraction arithmetic, English morphology, function learning), AToM systematically outperforms LLM-based and standard Bayesian teaching models. In human experiments, both AToM and LLMs outperform non-adaptive random example selection. Our results highlight both the difficulty of the adaptive teaching task and the potential of learned adaptive models for solving it.