AIDispute over the claim that 8% of The Pile is Christian
A user challenges a figure attributed to Tim Hwang, arguing that keyword-based classification counts long books with passing mentions of Christianity as Christian sources.
TLDR
A user disputes a claim they attribute to Tim Hwang that 8% of The Pile is Christian. They say his method classifies sources as Christian once they hit a keyword threshold, and claim 90% of those sources are long books that mention Christianity in passing—not Christian books.
Combined views
888
2 Sources, first seen 4h ago
Dispute over the claim that 8% of The Pile is Christian
A user challenges a figure attributed to Tim Hwang, arguing that keyword-based classification counts long books with passing mentions of Christianity as Christian sources.
TLDR
A user disputes a claim they attribute to Tim Hwang that 8% of The Pile is Christian. They say his method classifies sources as Christian once they hit a keyword threshold, and claim 90% of those sources are long books that mention Christianity in passing—not Christian books.