Hudson Golino
@hudsongolino.bsky.social
Associate Professor of Quantitative Methods at the Department of Psychology - University of Virginia. NLP/LLMs, Network Science, Info Theory, Psychometrics. Web: tinyurl.com/HudsonGolino
Web 2: https://r-ega.net
Web 2: https://r-ega.net
BREAKTHROUGH COMING NEXT WEEK! After months of very intensive work, we are about to release the results of our "Exploratory Graph Analysis Item Parameters" simulation. We've successfully derived IRT-compatible parameters directly from EGA! Stay tuned, folks!
May 9, 2025 at 2:58 PM
BREAKTHROUGH COMING NEXT WEEK! After months of very intensive work, we are about to release the results of our "Exploratory Graph Analysis Item Parameters" simulation. We've successfully derived IRT-compatible parameters directly from EGA! Stay tuned, folks!
Are you interested in hierarchical dimensionality analysis? Here's our new "Taxonomic Graph Analysis" used to model the IPIP-NEO Personality Hierarchy. The project is led by Andrew Samo and Alexander Christensen, with the collaboration of Luis Garrido, Paco Abad, Sam McAbee, and me! 1/n
April 22, 2025 at 12:30 PM
Are you interested in hierarchical dimensionality analysis? Here's our new "Taxonomic Graph Analysis" used to model the IPIP-NEO Personality Hierarchy. The project is led by Andrew Samo and Alexander Christensen, with the collaboration of Luis Garrido, Paco Abad, Sam McAbee, and me! 1/n
A new approach for LLM Interpretability: Mapping the Emergence of Semantic Structure in GPT-2. See the project details and initial findings below:
April 10, 2025 at 3:11 PM
A new approach for LLM Interpretability: Mapping the Emergence of Semantic Structure in GPT-2. See the project details and initial findings below:
What happens when we use Quantitative Psychology methods to understand and dissect Large Language Models? Follow me this week for a very interesting journey on the semantic dimensionality emergence and structural shifts in LLMs...
April 9, 2025 at 6:07 PM
What happens when we use Quantitative Psychology methods to understand and dissect Large Language Models? Follow me this week for a very interesting journey on the semantic dimensionality emergence and structural shifts in LLMs...
New paper alert: Revised Network Loadings, by Alexsander P. Christensen, Hudson Golino, Francisco Abad, and Luis Garrido (2025). Link: link.springer.com/article/10.3...
March 27, 2025 at 2:30 PM
New paper alert: Revised Network Loadings, by Alexsander P. Christensen, Hudson Golino, Francisco Abad, and Luis Garrido (2025). Link: link.springer.com/article/10.3...
The Generative Psychometrics and Exploratory Graph Analysis Lab has a new tag and a brand new room at UVa
March 6, 2025 at 8:10 PM
The Generative Psychometrics and Exploratory Graph Analysis Lab has a new tag and a brand new room at UVa
Come join me next week (March 12 and 13) for the very first: Automatic Item Generation and Validation: A Short Course, A Network Integrated Approach using Large Language Models in R. codehorizons.com/Seminars/aut...
March 3, 2025 at 5:49 PM
Come join me next week (March 12 and 13) for the very first: Automatic Item Generation and Validation: A Short Course, A Network Integrated Approach using Large Language Models in R. codehorizons.com/Seminars/aut...
We did it!!! A full-blown, 100% in-silica item development and validation (structural validity and item quality) for PERFORMANCE items using a new AI-GENIE developed exclusively for performance-based assessment/tests. Accuracy (NMI = .95). Mean Difference between DIFFICULT and EASY items: 3.34 1/n
February 14, 2025 at 2:59 PM
We did it!!! A full-blown, 100% in-silica item development and validation (structural validity and item quality) for PERFORMANCE items using a new AI-GENIE developed exclusively for performance-based assessment/tests. Accuracy (NMI = .95). Mean Difference between DIFFICULT and EASY items: 3.34 1/n
Our UVA Research Award Dinner last week, with our beloved friends Steve Boker and Valerie Larsen
February 10, 2025 at 4:49 PM
Our UVA Research Award Dinner last week, with our beloved friends Steve Boker and Valerie Larsen
Celebrating the 2024 Research Excellence Award at @UVA with my beautiful and brilliant wife, and two of the most inspiring people at UVa, our amazing president Jim Ryan and our great provost, Iam Baucom.
February 6, 2025 at 11:42 AM
Celebrating the 2024 Research Excellence Award at @UVA with my beautiful and brilliant wife, and two of the most inspiring people at UVa, our amazing president Jim Ryan and our great provost, Iam Baucom.
Coming to EGAnet soon: The Total Entropy Fit Index Bootstrap Test. Compare two dimensionality structures using bootstrapped TEFI values. Includes non-parametric tests and sweet visualization of TEFI distributions. Perfect for robust network comparison!
January 29, 2025 at 8:03 PM
Coming to EGAnet soon: The Total Entropy Fit Index Bootstrap Test. Compare two dimensionality structures using bootstrapped TEFI values. Includes non-parametric tests and sweet visualization of TEFI distributions. Perfect for robust network comparison!
Despite several tragedies going on (including a personal one), let me share some good news. I am very happy to receive the 2024 University of Virginia (@UVA) Research Excellence Award, together with my dear colleague Steve Boker (that received the Distinguished Researcher Award).
January 10, 2025 at 3:02 PM
Despite several tragedies going on (including a personal one), let me share some good news. I am very happy to receive the 2024 University of Virginia (@UVA) Research Excellence Award, together with my dear colleague Steve Boker (that received the Distinguished Researcher Award).
Results:
Comparing the In-Silica Structural Validity with the Structural Validity obtained in the Human Samples:
Comparing the In-Silica Structural Validity with the Structural Validity obtained in the Human Samples:
November 29, 2024 at 4:52 PM
Results:
Comparing the In-Silica Structural Validity with the Structural Validity obtained in the Human Samples:
Comparing the In-Silica Structural Validity with the Structural Validity obtained in the Human Samples:
EGA/BootEGA/Item Stability for the 5 Empirical (real human) Samples: 26/n
November 27, 2024 at 9:20 PM
EGA/BootEGA/Item Stability for the 5 Empirical (real human) Samples: 26/n
Participants provided consent, completed a survey, and concluded with a set of demographics questions (age, race/ethnicity, gender identity; Table 4). This study and subsequent surveys were all approved by the University’s institutional review board. 25/n
November 27, 2024 at 9:20 PM
Participants provided consent, completed a survey, and concluded with a set of demographics questions (age, race/ethnicity, gender identity; Table 4). This study and subsequent surveys were all approved by the University’s institutional review board. 25/n
Toy Example: 19/n
November 27, 2024 at 9:20 PM
Toy Example: 19/n
though even this relatively small change shows that the AI- GENIE is a sophisticated tool that is equipped to handle items generated by any of the five models at any temperature setting for any Big Five personality traits. 16/n
November 27, 2024 at 9:20 PM
though even this relatively small change shows that the AI- GENIE is a sophisticated tool that is equipped to handle items generated by any of the five models at any temperature setting for any Big Five personality traits. 16/n
AI-GENIE has the potential to save researchers months, if not years, of time and thousands, if not tens of thousands, of U.S. dollars. Similar research has already found evidence that language embeddings and/or language models can approximate 9/n
November 27, 2024 at 9:20 PM
AI-GENIE has the potential to save researchers months, if not years, of time and thousands, if not tens of thousands, of U.S. dollars. Similar research has already found evidence that language embeddings and/or language models can approximate 9/n
TEFI and NMI cannot be used separately for the optimization. The plot below shows that TEFI and NMI cannot be used in separate for the optimization.
November 25, 2024 at 5:55 PM
TEFI and NMI cannot be used separately for the optimization. The plot below shows that TEFI and NMI cannot be used in separate for the optimization.
TEFI per number of embedded dimensions (from 1 to 260, each entry representing positions from 3 to 1500 in the LLM embedding space).
November 25, 2024 at 5:55 PM
TEFI per number of embedded dimensions (from 1 to 260, each entry representing positions from 3 to 1500 in the LLM embedding space).