PulseAugur
LIVE 13:08:44
research · [2 sources] ·
0
research

SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs

Researchers have introduced SHAPE, a new benchmark designed to evaluate the safety, helpfulness, and pedagogical effectiveness of educational Large Language Models (LLMs). The benchmark addresses a vulnerability known as "pedagogical jailbreaks," where students attempt to elicit direct answers rather than guided learning. SHAPE includes over 9,000 student-question pairs and a proposed graph-augmented tutoring pipeline to improve LLM performance in educational settings. AI

Summary written by None from 2 sources. How we write summaries →

IMPACT Introduces a new benchmark and evaluation framework for educational LLMs, potentially improving their safety and pedagogical approach.

RANK_REASON The cluster describes an academic paper introducing a new benchmark and methodology for evaluating educational LLMs.

Read on arXiv cs.CL →

COVERAGE [2]

  1. arXiv cs.CL TIER_1 · Sihang (Nagi), Zhao, Kangrui Yu, Youliang Yuan, Pinjia He, Hongyi Wen ·

    SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs

    arXiv:2604.22134v1 Announce Type: new Abstract: Large Language Models (LLMs) have been widely explored in educational scenarios. We identify a critical vulnerability in current educational LLMs, pedagogical jailbreaks, where students use answer-inducing prompts to elicit solution…

  2. arXiv cs.CL TIER_1 · Hongyi Wen ·

    SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs

    Large Language Models (LLMs) have been widely explored in educational scenarios. We identify a critical vulnerability in current educational LLMs, pedagogical jailbreaks, where students use answer-inducing prompts to elicit solutions rather than scaffolded instructions. To enable…