Experimental browser for the Atmosphere
Loading post...
{ "uri": "at://did:plc:fbkipu3c5tczzv7qack542zo/app.bsky.feed.like/3llrvovnimk2b", "cid": "bafyreib4e6gx6vhaxgsbywqgmiw3gc25ddejkztcu2u3uoyajw7xojcrpi", "value": { "$type": "app.bsky.feed.like", "subject": { "cid": "bafyreigtut2ncutmawlv2redp5rj7lrywuyjny2wbo5g3fb6vjhyts5tzi", "uri": "at://did:plc:gatrjx7cdtlh3kbhsq4qltjk/app.bsky.feed.post/3llroejjics25" }, "createdAt": "2025-04-01T22:18:14.848Z" } }
Notably, we show DeepSeek-R1 has a ‘sweet spot’ of reasoning, where extra inference time can impair model performance and continuously scaling length of thoughts does not necessarily increase performance.
Apr 1, 2025, 8:06 PM