ATProto Browser

Record data

{
  "uri": "at://did:plc:5ugfdqftlt7isp4lvn5hykb2/app.bsky.feed.post/3lk23gy6bd22p",
  "cid": "bafyreibruexiplkhedteqeygtiulhx6zexuny5qsjryw7ypfpv5gpe7epe",
  "value": {
    "text": "Do LLMs understand probability distributions? Can they serve as effective simulators of probability? No!\nHowever, in our latest paper that via in-context learning, LLMs update their broken priors in a manner akin to Bayseian updating.\n\n📝 arxiv.org/abs/2503.04722",
    "$type": "app.bsky.feed.post",
    "embed": {
      "$type": "app.bsky.embed.external",
      "external": {
        "uri": "https://arxiv.org/abs/2503.04722",
        "thumb": {
          "$type": "blob",
          "ref": {
            "$link": "bafkreifybtmgi5p3pxui3jso5onrtykvylqrbt6lrralsflgppistcbdie"
          },
          "mimeType": "image/jpeg",
          "size": 178482
        },
        "title": "Enough Coin Flips Can Make LLMs Act Bayesian",
        "description": "Large language models (LLMs) exhibit the ability to generalize given few-shot examples in their input prompt, an emergent capability known as in-context learning (ICL). We investigate whether LLMs uti..."
      }
    },
    "langs": [
      "en"
    ],
    "facets": [
      {
        "index": {
          "byteEnd": 265,
          "byteStart": 241
        },
        "features": [
          {
            "uri": "https://arxiv.org/abs/2503.04722",
            "$type": "app.bsky.richtext.facet#link"
          }
        ]
      }
    ],
    "createdAt": "2025-03-10T17:32:06.388Z"
  }
}