ATProto Browser

Experimental browser for the Atmosphere

Post

The other thing I've seen in the literature is simply prepending this information to the model, but that's problematic if the information is only available at training time and needs to be inferred at testing time. I could maybe train an auxiliary model to predict the right value for the token?

Dec 13, 2024, 5:59 PM

Loading post...

Record data

{
  "uri": "at://did:plc:uuvoat235entezwm4564e27m/app.bsky.feed.like/3ldbcg3c3lp2e",
  "cid": "bafyreied4l4wa2o6llkggkumahlygnqrxcqnr534ffbudihqe76ketsbay",
  "value": {
    "$type": "app.bsky.feed.like",
    "subject": {
      "cid": "bafyreibir3nfpfyuhbdbnmig6ebbg2wxxeiecoqi7zbzpz3nsqtlqypgme",
      "uri": "at://did:plc:mray7klv7leelnlkynuk57up/app.bsky.feed.post/3ld7efvfqvs2q"
    },
    "createdAt": "2024-12-14T12:29:08.900Z"
  }
}