4 ms·
Right, an ML model failing on a task it wasn't trained on is not news. That's not to say there aren't problems LLMs can do surprisingly well on even though they
by morkalork 1y ago
Right, an ML model failing on a task it wasn't trained on is not news. That's not to say there aren't problems LLMs can do surprisingly well on even though they weren't explicitly trained to do. Those cases are certainly newsworthy. But that's not the case here.