README: tighten Quick Start 3 phrasing

v2.4: README at sample-subdirectory level explaining how the representative sample was created (per NeurIPS E&D Dataset Large URL guideline)

v2.3: croissant url -> HF dataset page; README adds 'Representative sample for review' section per NeurIPS E&D > 4 GB requirement

v2.2: croissant — add rai:dataBiases, rai:personalSensitiveInformation, rai:hasSyntheticData, prov:wasDerivedFrom, prov:wasGeneratedBy; expand limitations / use-cases / social-impact

Add files using upload-large-folder tool

Add files using upload-large-folder tool

v2.1: README — fix splits/probe paths, add L2-norm note, add reproduction quick-start with balanced-subsample pattern

v2: refresh README + croissant.json for expanded artifact set

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Upload croissant.json with huggingface_hub

Upload folder using huggingface_hub

initial commit