3 ms·
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
by syntaxing 11mo ago
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
- pehtis 11mo agoSpecifically for GPT-OSS I had great success with this: https://old.reddit.com/r/LocalLLaMA/comments/1ng9dkx/gptoss_jailbreak_system_prompt/ https://old.reddit.com/r/LocalLLaMA/comments/1ng9dkx/gptoss_...