New Local LLM Grading Tool for STEM Subjects Unveiled
Researchers have developed LaTA, a locally hosted LLM-based automated grading system for engineering and science courses, addressing data privacy concerns while reducing assessment workloads.
What happened?
Researchers have presented LaTA (LaTeX Teaching Assistant), a new automated grading tool that employs a locally hosted Large Language Model (LLM) to assess student assignments in science and engineering. Designed for LaTeX-based submissions, the system follows a four-step process: ingestion, segmentation, assessment, and reporting. The locally hosted LLM, gpt-oss:120b, compares student work against a reference solution and applies a YAML-based grading rubric with binary scoring per item.
Key facts
| Verktygets namn | LaTA (LaTeX Teaching Assistant) |
|---|---|
| LLM-modell | gpt-oss:120b |
| Implementeringsplats | Oregon State University, ME 373 |
| Implementeringstid | Vintern 2026 |
| Målgrupp | Högre ingenjörs- och naturvetenskapskurser |
”Large-language-model (LLM) graders promise to relieve the grading burden of upper-division STEM courses, but most deployments to date send student work to third-party APIs, violating FERPA and exposing institutions to data risk while requiring substantial assignment modification.”
Why it matters
LaTA aims to address data privacy issues arising when student data is transmitted to third-party LLM services, which often violates regulations such as FERPA in the United States. By running the LLM locally, higher education institutions can maintain control over student data and mitigate risks of information leakage. The tool has the potential to streamline the assessment of assignments in advanced engineering and science courses—a process traditionally time-consuming for instructors.
Who is affected?
LaTA is primarily aimed at universities and colleges offering engineering and science programmes, as well as instructors and lecturers in these fields. Students are affected as their assignments may be graded by AI, though with enhanced data privacy. Developers and researchers in AI and educational technology may find interest in the open-source architecture used for local LLM deployment.
What else you should know
LaTA was tested during the winter of 2026 in course ME 373 (Mechanical Engineering Methods) at Oregon State University, where it was used to grade weekly assignments. The researchers have not specified the exact date for this "winter 2026", which may suggest a future implementation or an academic projection.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Vilka bolag berörs?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
AI-verktyg i artikeln
Topics
Get similar news straight to your inbox
The reader's room
Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.
Sign in to submit a comment or question.
Read the article through your role
- Decide whether this affects strategy over 6–12 months or is just noise.
- Discuss with leadership: do we own the right question or does ownership need to move?
- Ask: what risk are we taking by NOT acting on this this quarter?
Generated angle — not editorial analysis of "New Local LLM Grading Tool for STEM Subjects Unveiled"