2026-01-14 22:27 UTC
🧩📚 How can we #evaluate poetic interpretations produced by #LLMs in a way that aligns with #LiteraryStudies?
The authors propose a #workflow grounded in reference #interpretations and argumentative structures from literary scholarship, enabling the #assessment of LLM-generated readings.
Replies (1)
-
@jcls@fedihum.org 2026-01-14 22:31
🔍🧠 Their experiments show that #LLMs can produce reasonable poem descriptions, but struggle with more abstract interpretion, highlighting where #NLG currently meets its #limits in #LiteraryInterpretation. #LiteraryComputing #Evaluation #GenerativeModels