Comment by syntaxing

Amazing. I’m eager to see what the results for GPT-OSS is like. It’s a great model but the “safety alignment” ruins it