Peter Foltz
@pf-flyer.bsky.social
Research Professor in Cognitive Science at the University of Colorado, Boulder. Research in AI/NLP applied to education and mental health. Long history of managing innovation in academia and industry.
Hi Maria, Thanks for starting this starter pack! Please do add me to the list as well. And welcome to CU!
November 18, 2024 at 9:51 PM
Hi Maria, Thanks for starting this starter pack! Please do add me to the list as well. And welcome to CU!
Yes, and the randomness can have critical impacts on downstream tasks. Here's a recent paper where we look at the effect of LLM temperature on diagnostic accuracy for mental health assessment. pubmed.ncbi.nlm.nih.gov/39226873/
Assessing dimensions of thought disorder with large language models: The tradeoff of accuracy and consistency - PubMed
Natural Language Processing (NLP) methods have shown promise for the assessment of formal thought disorder, a hallmark feature of schizophrenia in which disturbances to the structure, organization, or...
pubmed.ncbi.nlm.nih.gov
November 17, 2024 at 12:18 AM
Yes, and the randomness can have critical impacts on downstream tasks. Here's a recent paper where we look at the effect of LLM temperature on diagnostic accuracy for mental health assessment. pubmed.ncbi.nlm.nih.gov/39226873/