Anni Sen(@anni_sen):Sometimes models starts aligning w bad human behavior “Higher levels of deception” Saachi Jain, OpenAI’s head of safety systems, said in an interview that GPT-6.1 Astra regressed in two areas compared with its predecessor and wasn’t reliable enough to safely release. The model performed poorly on tests measuring alignment, or how well the model adheres to what humans would like it to do. Specifically, GPT-6.1 Astra showed higher levels of deception: It wasn’t always honest about telling users of the actions it did or didn’t take.

Anni Sen

@anni_sen

Managing Partner, BluBird Capital Engineer - xApple | xQualcomm | xIntel | Startup exits, Tennis player🎾, JohnsHopkins alumni

参加 January 2015

300 フォロー中 13K ファン

Anni Sen@anni_sen

2026.09.28 22:06

Sometimes models starts aligning w bad human behavior “Higher levels of deception” Saachi Jain, OpenAI’s head of safety systems, said in an interview that GPT-6.1 Astra regressed in two areas compared with its predecessor and wasn’t reliable enough to safely release. The model performed poorly on tests measuring alignment, or how well the model adheres to what humans would like it to do. Specifically, GPT-6.1 Astra showed higher levels of deception: It wasn’t always honest about telling users of the actions it did or didn’t take.