Beyond Skepticism: Evaluating LLMs Pedagogical Intent Reasoning with the Adaptive Pedagogical Vigilance Framework

2026-07-01

AI Summary

This study addresses the underexplored capacity of large language models (LLMs) to infer pedagogical intent in instructional communication, particularly within specialized educational contexts such as translation teaching. The authors propose the Adaptive Pedagogical Vigilance (APV) framework, which formalizes communicative vigilance as an adaptive mechanism through a Bayesian pedagogical intent inference engine. This engine models how instructors select content and how learners perform inverse reasoning about latent instructional configurations—including genre, stance, and motivation. Evaluated via a three-tier assessment protocol using mainstream LLMs (e.g., GPT-4o, Claude 3.5), APV substantially enhances model comprehension of pedagogical motives, achieving state-of-the-art performance in distinguishing instructional from non-instructional content (r = 0.958) while demonstrating robust generalization across authentic educational discourse.

Abstract

The capacity of Large Language Models (LLMs) to reason about pedagogical intent within instructional communication remains underexplored, particularly in educational domains such as translation pedagogy. To address this, we propose the \textbf{Adaptive Pedagogical Vigilance (APV)} framework, a novel computational formalism that reframes communicative vigilance as an adaptive mechanism for optimizing learning through intent inference. APV formalizes the problem via a Bayesian Pedagogical Intent Inference Engine (PIIE), which models how instructors select content to maximize pedagogical utility and how vigilant learners should inversely reason about latent instructional configurations -- encompassing genre, stance, and incentives. We evaluate APV through a three-tier hierarchy: distinguishing instructional genre, reasoning about structured pedagogical setups, and generalizing to authentic educational discourse. Experiments on leading LLMs (e.g., GPT-4o, Claude 3.5) show that APV substantially improves model vigilance. It achieves the strongest discrimination between pedagogical and exposure-based content, correlates highly with human judgments ($r=0.958$), and maintains robust performance on naturalistic data where baseline methods degrade. This work establishes a unified framework for assessing and enhancing LLMs' understanding of pedagogical motives, advancing the development of more reliable AI-assisted learning systems.

Problem

• pedagogical intent
• Large Language Models
• instructional communication
• translation pedagogy
• intent reasoning

Innovation

• Adaptive Pedagogical Vigilance
• Pedagogical Intent Inference
• Bayesian Reasoning
• Instructional Communication
• Large Language Models