LLM-generated podcast previews beat a feature-engineered multi-model baseline in human evaluation and production A/B testing while processing episodes five times faster.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Transforming Podcast Preview Generation: From Expert Models to LLM-Based Systems
LLM-generated podcast previews beat a feature-engineered multi-model baseline in human evaluation and production A/B testing while processing episodes five times faster.