Experimental browser for the Atmosphere
Loading post...
{ "uri": "at://did:plc:t7jbj4w3uo2sus4bc3znsspa/app.bsky.feed.like/3lk752vyay52x", "cid": "bafyreicteixyzsimk2aqpr6r6orkabe7xjpioru3myj3nmpoksxhjcxzwq", "value": { "$type": "app.bsky.feed.like", "subject": { "cid": "bafyreidcddkjtelax7wicpskwfe7uotz3w7wxdd7ng4c2czsi72p65hfj4", "uri": "at://did:plc:gjh7vfpynd4diwhpjnljqgwn/app.bsky.feed.post/3lk6scfnppc2n" }, "createdAt": "2025-03-12T17:44:27.587Z" } }
New preprint w/ @jennhu.bsky.social @kmahowald.bsky.social : Can LLMs introspect about their knowledge of language? Across models and domains, we did not find evidence that LLMs have privileged access to their own predictions. 🧵(1/8)
Mar 12, 2025, 2:31 PM