//
sign in
Post
by @danabra.mov
PostEmbed
by @danabra.mov
Record
by @jimpick.com
Record
by @atsui.org
+ new component
Post
3/ To do this at scale we built NarraBert, a RoBERTa-based model validated against human and LLM annotations, and released NarraDolma: ~2.9M passages across ~785K Dolma documents, each scored on all 12 narrative dimensions.