A DeBERTa grader trained on 322,538 GPT-3.5-generated question-answer pairs scored high on synthetic tests and outperformed GPT-4o on a small real student dataset, but the evidence has significant caveats.
Applications and Rules
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Simulation as Reality? The Effectiveness of LLM-Generated Data in Open-ended Question Assessment
A DeBERTa grader trained on 322,538 GPT-3.5-generated question-answer pairs scored high on synthetic tests and outperformed GPT-4o on a small real student dataset, but the evidence has significant caveats.