Record🤔

uri:

"at://did:plc:26hkfmtr7jrchexamh7hk6rz/app.bsky.feed.post/3ks5gu565es2r"

cid:

"bafyreiazo2s3gaaa4fknmclk6sndepdmsfhlgwxqztzxp535zrgsokhtqu"

value:

text:

"We evaluated 14 models trained on 400M & 2B samples on classification. As training data increased, so did probability of misclassifying human images as offensive class such as criminal
Paper: https://arxiv.org/abs/2405.04623
Code: https://github.com/SepehrDehdashtian/the-dark-side-of-dataset-scaling"

$type:

"app.bsky.feed.post"

embed:

$type:

"app.bsky.embed.images"

images:

• alt:

  ""

  image:

  

  View blob content

  $type:

  "blob"

  ref:

  $link:

  "bafkreidejro5jblgmwkszcshojhhxoofm5ljkddmwmc4jiamuze2tegmce"

  mimeType:

  "image/jpeg"

  size:

  785116

facets:

• index:

  byteEnd:

  224

  byteStart:

  192

  features:

  • uri:

    "https://arxiv.org/abs/2405.04623"

    $type:

    "app.bsky.richtext.facet#link"

• index:

  byteEnd:

  300

  byteStart:

  231

  features:

  • uri:

    "https://github.com/SepehrDehdashtian/the-dark-side-of-dataset-scaling"

    $type:

    "app.bsky.richtext.facet#link"

createdAt:

"2024-05-10T15:29:36.692Z"