2 ms·
What is the evidence for 1) ? I thought that the latest models were getting "somewhere" with fairly trivial reasoning tests like ARC-1
by sgt101 2y ago
What is the evidence for 1) ? I thought that the latest models were getting "somewhere" with fairly trivial reasoning tests like ARC-1
- sealeck 2y agoIt may be that you can just find the solution for these tests by interpolating from a very large dataset.