How Can I Improve? Using GPT to Highlight the Desired and Undesired Parts of Open-ended Responses
arXiv:2405.00291 · doi:10.1051/0004-6361/202349120
Abstract
Automated explanatory feedback systems play a crucial role in facilitating learning for a large cohort of learners by offering feedback that incorporates explanations, significantly enhancing the learning process. However, delivering such explanatory feedback in real-time poses challenges, particularly when high classification accuracy for domain-specific, nuanced responses is essential. Our study leverages the capabilities of large language models, specifically Generative Pre-Trained Transformers (GPT), to explore a sequence labeling approach focused on identifying components of desired and less desired praise for providing explanatory feedback within a tutor training dataset. Our aim is to equip tutors with actionable, explanatory feedback during online training lessons. To investigate the potential of GPT models for providing the explanatory feedback, we employed two commonly-used approaches: prompting and fine-tuning. To quantify the quality of highlighted praise components identified by GPT models, we introduced a Modified Intersection over Union (M-IoU) score. Our findings demonstrate that: (1) the M-IoU score effectively correlates with human judgment in evaluating sequence quality; (2) using two-shot prompting on GPT-3.5 resulted in decent performance in recognizing effort-based (M-IoU of 0.46) and outcome-based praise (M-IoU of 0.68); and (3) our optimally fine-tuned GPT-3.5 model achieved M-IoU scores of 0.64 for effort-based praise and 0.84 for outcome-based praise, aligning with the satisfaction levels evaluated by human coders. Our results show promise for using GPT models to provide feedback that focuses on specific elements in their open-ended responses that are desirable or could use improvement.
arXiv comment11 pages, full research paper, EDM 2024
References in corpus (10)
- Observations of the Icy Universe
- An Ice Age JWST inventory of dense molecular cloud ices
- H-atom addition and abstraction reactions in mixed CO, H2CO and CH3OH ices: an extended view on complex organic molecule formation
- CRDB: a database of charged cosmic rays
- The c2d Spitzer Spectroscopic Survey of Ices Around Low-Mass Young Stellar Objects. IV. NH3 and CH3OH
- Spitzer Mid-infrared Spectroscopy of Ices Toward Extincted Background Stars
- Interstellar dust charging in dense molecular clouds: cosmic ray effects
- Ices in the Quiescent IC 5146 Dense Cloud
- Production of atomic hydrogen by cosmic rays in dark clouds
- Linking ice and gas in the Serpens low-mass star-forming region