Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Podcast

Benchmarks

Task NameDataset NameSOTA ResultTrend
Diversitypodcast
Win Count56
6
Comprehensivenesspodcast
Win Count50
6
GraphRAG Content Quality Evaluationpodcast Leiden level C3 post-cutoff
Comprehensiveness Win50
6
GraphRAG Content Quality Evaluationpodcast Leiden level C2 post-cutoff
Comprehensiveness Win Count50
6
Directnesspodcast C3 community level post-cutoff (test)
Win Count64
6
Directnesspodcast C2 community level post-cutoff (test)
Wins62
6
Diversitypodcast Leiden level C1
Win Rate72
6
Diversitypodcast Leiden level C0
Win Rate68
6
Comprehensivenesspodcast Leiden level C1
Win Rate (C1 Podcast)68
6
Comprehensivenesspodcast Leiden level C0
Win Rate64
6
Community Summary Evaluationpodcast C3 (post-cutoff)
Comprehensiveness Win53
6
Community Summary Evaluationpodcast C2 (post-cutoff)
Comprehensiveness Win58
6
Speaker-Attributed Automatic Speech RecognitionPodcast (test)
CER4.46
4
Showing 13 of 13 rows