jevPOST
Jev 被问「你是谁」时自称 OpenAI
作者用离散选择题问 Jev 的自我认知，Jev 答自己是 OpenAI；他确认团队从未教模型说自己是 typesafe/jev。
Diogo Almeida
@CompleteSkeptic2026.09.20收录原推发布于 2026.09.19
Diogo Almeida @CompleteSkeptic原推发布于 2026.09.19
@suchenzang I'd be curious about the choices! when I ask it in a discrete choice, it says it's OpenAI (which would be calibrated for the internet distribution of generated text) we definitely do not teach the model that it is typesafe/jev
x.com/completeskeptic/status/2101141848488333411采集于 2026.09.20 10:16