Table of Links
Abstract and 1 Introduction
2. Background
2.1 Effective Tutoring Practice
2.2 Feedback for Tutor Training
2.3 Sequence Labeling for Feedback Generation
2.4 Large Language Models in Education
3. Method
3.1 Dataset and 3.2 Sequence Labeling
3.3 GPT Facilitated Sequence Labeling
3.4 Metrics
4. Results
4.1 Results on RQ1
4.2 Results on RQ2
5. Discussion
6. Limitation and Future Works
7. Conclusion
8. Acknowledgments
9. References
APPENDIX
A. Lesson Principles
B. Input for Fine-Tunning GPT-3.5
C. Scatter Matric of the Correlation on the Outcome-based Praise
D. Detailed Results of Fine-Tuned GPT-3.5 Model’s Performance
A. LESSON PRINCIPLES
The following is the principle that a correct response should follow:
Praising students for working hard and putting forth effort is a great way to increase student motivation. When the learning gets tough, giving correct praise is a powerful strategy to encourage students to keep going.
The correct response should be :
-perceived as sincere, earned, and truthful.
-specific by giving details of what the student did well.
-immediate with praise given right after the student action.
-authentic and is not repeated often, such as “great job” which loses meaning and becomes predictable.
-focused on the learning process, not ability (AJTutoring.com, 2022)
Correct responses must follow some, but not all the above. There are two types of praise responses: Effort and Outcome praise
-Effort praise focuses on the learning process. Effort praise recognizes students for putting forth effort and persevering through the learning process instead of focusing on whether a student got the problem correct or pure ability.
-Outcome praise showcases student’s achievements, such as getting a grade A on an assignment or getting a problem correct, and is often, but not always, associated with unproductive praise.
To receive full credit of correct praise, tutors cannot just say ”great job” and praise with no specific reasoning. Tutors need to praise for effort AND be positive and encouraging.
