Figure 5 - uploaded by Francisco Olivos
Content may be subject to copyright.
Download
View publication
Standard deviation of 100 scoring tasks without prompting template for each setting.

Standard deviation of 100 scoring tasks without prompting template for each setting.

Source publication
Figure 3. Scoring prompt utilizing GPT models' sociological knowledge.
Figure 4. Scoring prompt providing template answers.
Figure 5. Standard deviation of 100 scoring tasks without prompting...
Figure 6. Correlation between total human and GPT scoring without...
+4 Figure 7. Bland-Altman graph for average scores and differences between...
Assessing instructor-AI cooperation for grading essay-type questions in an introductory sociology course
Preprint
Full-text available
  • Jan 2025
This study explores the use of artificial intelligence (AI) as a complementary tool for grading essay-type questions in higher education, focusing on its consistency with human grading and potential to reduce biases. Using 70 handwritten exams from an introductory sociology course, we evaluated generative pre-trained transformers (GPT) models' perf...
Cite
Download full-text

Contexts in source publication

Context 1
... temperatures generate more deterministic results; therefore, we may expect lower variability in the outputs produced by settings at temperature 0.2. Figure 5 displays the standard deviation across 100 runs of the same task under different settings of model and temperature for the GPT models. ...
Context 2
... Figure 5 about here] ...

Similar publications

Figure 1. Graphical representation of the MSE values using artificial data.
Mean squared error (MSE) values using the simulated data.
Percent relative efficiency (PRE) values using the simulated data.
Mean absolute percentage error (MAPE) values using the simulated data.
+1 Cont.
A Comparative Study of New Ratio-Type Family of Estimators Under Stratified Two-Phase Sampling
Article
Full-text available
  • Jan 2025
Two-phase sampling is a useful technique for sample surveys, particularly when prior auxiliary data is not accessible. The ranks of the auxiliary variable often coincide with those of the research variable when two variables are correlated. By considering this relationship, we can significantly increase estimator accuracy. In this paper, we use the...
View