CONVERSATIONAL RESEARCH TOOLSStARCASM & DANG
Explore how two trained language models classify text and support short conversational exchanges.
DANG
Try text analysis, explore the model response, and submit corrections for review.
Open DANG
StARCASM
Open the companion research demonstration.
Companion bot link will appear after deployment.
Before trying the demo
A reviewer login is required. Use fictional examples and do not enter names, medical records, or other personal information. Submitted corrections are saved for review. Chat text may be sent to the configured external response provider; classification runs on the hosted model.
These tools classify language. The scores do not measure cognitive ability, establish a diagnosis, or demonstrate rehabilitation benefit.