Concurrent Session #1: Harnessing Large Language Models for Automatic Grading and Hint Generation
Alternative Title
Harnessing Large Language Models for Automatic Grading and Hint Generation
Contributor
University of Central Florida. Faculty Center for Teaching and Learning; University of Central Florida. Division of Digital Learning; Teaching and Learning with AI Conference (2024 : Orlando, Fla.)
Location
Seminole A
Start Date
22-7-2024 1:00 PM
End Date
22-7-2024 1:30 PM
Publisher
University of Central Florida Libraries
Keywords:
Automatic grading; Hint generation; AI analytics; Guidance rubrics; Student performance
Subjects
Grading and marking (Students)--Computer-assisted instruction; Grading and marking (Students)--Computer programs; Artificial intelligence--Computer-assisted instruction; Grading and marking (Students)--Data processing; Grading and marking (Students)--Evaluation
Description
Due to hallucinations and inconsistencies, using ChatGPT directly for grading assignments typically yields poor results. However, by employing sequences of specific prompts and detailed “guidance rubrics,” we can enhance the grading quality of AI. Instructors can also request compiled, detailed analytics that pinpoint common student errors and track improvements over time. These analytics help instructors improve guidance rubrics and steer the development of learning material.
Language
eng
Type
Presentation
Rights Statement
All Rights Reserved
Audience
Faculty, Instructional designers, Educators
Recommended Citation
Bohacek, Stephan, "Concurrent Session #1: Harnessing Large Language Models for Automatic Grading and Hint Generation" (2024). Teaching and Learning with AI Conference Presentations. 3.
https://stars.library.ucf.edu/teachwithai/2024/monday/3
Concurrent Session #1: Harnessing Large Language Models for Automatic Grading and Hint Generation
Seminole A
Due to hallucinations and inconsistencies, using ChatGPT directly for grading assignments typically yields poor results. However, by employing sequences of specific prompts and detailed “guidance rubrics,” we can enhance the grading quality of AI. Instructors can also request compiled, detailed analytics that pinpoint common student errors and track improvements over time. These analytics help instructors improve guidance rubrics and steer the development of learning material.