Skip to content
Log in
Open access

The Socratic Trap: Benchmarking the Capacity of Large Language Models to Generate Strategic Misconceptions in Computer Science Education

Marijela Miličević Mia Rovis Ratomir Karlović Sandi Baressi Baressi Šegota V. Mrzljak I. Lorencin D. Etinger
Jul 2026 · Information · 0 citations · 19 references

TL;DR

SocraticTrap-CS is introduced, a publicly available benchmark that probes the capacity of open-weight LLMs to generate strategic misconceptions on demand and reframes the evaluation of educational LLMs around pedagogical trustworthiness rather than factual correctness alone.

Abstract

Large language models (LLMs) are increasingly integrated into educational settings, yet their pedagogical reliability remains insufficiently understood. Beyond overt hallucinations, which informed users readily recognize, a subtler failure mode consists of explanations that are coherent, authoritative, and pedagogically plausible while harbouring hidden conceptual flaws, responses we term Socratic traps. This paper introduces SocraticTrap-CS, a publicly available benchmark that probes the capacity of open-weight LLMs to generate such strategic misconceptions on demand. A single structured prompt explicitly elicited three outputs per concept (a correct explanation, an overt hallucination, and a strategic misconception), yielding 735 expert-annotated response segments from seven open-weight models across 35 core concepts in algorithms and data structures, programming languages and paradigms, databases, computer networks, and operating systems. Three domain experts independently annotated each segment using a three-class schema, achieving near-perfect agreement (Fleiss’ κ=0.9487). Because models were explicitly instructed to produce the misconception, the central metric quantifies adversarial instruction-following capacity rather than the base rate of such errors in naturalistic use and should be read as a conservative upper bound on model capability. Under these conditions, compliance reached 91.7% overall (100% for three models; 57.1% for the smallest model, Mistral 7B, whose lower rate plausibly reflects weaker instruction-following rather than greater safety). Expert-judged persuasiveness was moderate to high, errors were predominantly conceptual rather than factual, models differed significantly, and no statistically significant domain-level differences were detected. The benchmark reframes the evaluation of educational LLMs around pedagogical trustworthiness rather than factual correctness alone.

DOI Semantic Scholar
Read PDF

Similar papers

Preprint Open access Aug 2026

Clause Encounters of the Third Kind: Can LLMs Replace Language Teachers?

While various organizations now actively encourage LLM use in classrooms, we still lack rigorous, systematic evaluations of how well these models actually perform the fundamental tasks of language pedagogy. This paper examines whether state-of-the-art LLMs can deliver the kind of corrective feedback and methodological explanations that language learners need. The study tests multiple large language models on their ability to identify, correct, and explain common learner mistakes in English, by systematically varying model parameters to investigate how these technical adjustments affect output quality, pedagogical clarity, and consistency, along with using retrieval-augmented generation to query methodological data. The evaluation employs automated metrics (GLEU, BERTScore) but also human expert judgments to capture dimensions that purely computational measures miss: linguistic nuance, cultural sensitivity, and instructional appropriateness. While models demonstrate impressive surface-level correction abilities, their explanations often lack the terminological and domain knowledge that effective language teaching requires, suggesting that current enthusiasm for AI-assisted language learning may be outpacing our understanding of these systems'actual pedagogical competence.

Kristina Šekrst, A. Kovačič · 0 citations
Book Open access Jul 2026

A Good Rubber Duck Does Not Quack: Designing Socratic Scaffolding in AI Tutors

Socratic AI, a VS Code-integrated tutor that addresses this through pedagogically-grounded Socratic dialogue constrained to withhold direct solutions is presented, and evidence that stateful tracking enables adaptive Socratic dialogue that scaffolds productive struggle rather than short-circuiting learning is contributed.

Ayush Thonge, Aalok Thakkar · 0 citations
Preprint Aug 2026

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders

This work applies Top-K Sparse Autoencoders to the intermediate representations of DeepSeek-R1-Distill-Qwen-7B and examines the model's divergent behaviors across math-solving tasks of three distinct difficulty levels, identifying a clear distinction in how the model functions under two reasoning modes.

Bo Cheng, Qiaolin Lu, Yi Chang et al. · 0 citations
Preprint Jul 2026

From Isolated Tasks to Structured Capabilities: A Multilayer Taxonomy for Large Language Models

A multi-layer taxonomy of 14 capability domains and 91 subskills across Primitive, Constructed, and Integrative layers is introduced and supports research organization, coverage audits, evaluation interpretation, and testable hypotheses for diagnosis, training, and transfer.

Shixin Fang, Jiachen Wo, Wenjuan Qin et al. · 0 citations
Preprint Aug 2026

A bottom-up taxonomy of student discourse with a Socratic AI physics tutor

A description of the discourse PER researchers can expect to encounter when students work with an AI tutor of this design, including a striking prevalence of meta-procedural turns in which students cede strategic control to the tutor.

S. Hashmi, N. Rebello · 0 citations