研究者测试 abliterated GLM 5.3 后称其要么过度自信要么人类要完
AI summary
研究者 Matthew Green 表示，他获得了 abliterated GLM 5.3 的访问权限，并称该模型要么过度自信，要么意味着“我们全完了”。
Full text · Original
So I got access to abliterated GLM 5.3 and either it’s overconfident or we’re all screwed.
Translation in progress Some entries still include text in the original language.