general family · filed under data curation
High-recall web-data curation
A web-data pipeline shifted toward higher recall to preserve diversity while maintaining quality.
Also called high-recall web data.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
We shifted from a predominantly high-precision web pipeline toward a substantially higher-recall pipeline designed to preserve diversity while maintaining quality.
useddata curationin Laguna XS.2Poolside
Filed alongside
Other methods under data curation :: data sourcing.
Expert co-created training dataGitHub CrawlAgentic knowledge graph constructionCommon CrawlEssentialWebFinePDFsImage-code pairs and computer-use trajectoriesLong-document curationNemotron-3-Ultra corpusNemotron-CCNemotron-Post-Training-v3Star-threshold-filtered GitHub repositoriesStreaming training-data ingestionVideo frame extraction at 1 FPSVideo sampling parametersWeb knowledge graph construction and question generation