A pipeline converts CANDU textbook chapters into synthetic QA pairs using LLMs, embedding clustering, and similarity metrics, with no downstream validation.
They can revolutionize information retrieval, knowledge management, and even decision support systems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Unlocking the Potential of Large Language Models in the Nuclear Industry with Synthetic Data
A pipeline converts CANDU textbook chapters into synthetic QA pairs using LLMs, embedding clustering, and similarity metrics, with no downstream validation.