LaTA autograder uses local LLM to grade STEM coursework compliantly

Researchers have developed LaTA, an open-source autograder that uses a local LLM to grade STEM coursework without sending student data to third-party APIs. This FERPA-compliant system runs on commodity hardware and integrates with existing LaTeX workflows, grading assignments in minutes. Initial deployment at Oregon State University showed a low error rate and led to improved student performance and confidence. AI

IMPACT Provides a FERPA-compliant, on-premises LLM grading solution that could reduce data risks for educational institutions and improve student outcomes.

RANK_REASON Academic paper detailing a new open-source autograder system for educational coursework. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• AGPLv3
• FERPA
• gpt-oss:120b
• LaTeX
• LLM
• Oregon State University

• paper
• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →