Work

PLAR: Prompt-Level Alignment of Student Reasoning with Teacher Traces

Year: 2026

Type: other

Abstract: Code and per-run artifacts for *Prompt-Level Alignment of Student Reasoning with Teacher Traces*. The method optimizes a student model's **prompt** so that its reasoning agrees with a largerteacher's, using an LLM judge as the optimization objective. The student's weights are never touched.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.21964986

Author: Koralewski, Marcin

Language: English

Topic: Intelligent Tutoring Systems and Adaptive Learning

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Quality education

Open Access status: green

OpenAlex ID: W7203575520

Cites:

Cited by:

Related to: 10