sabreW4K3@lazysoci.al to Tech@programming.devEnglish · 2 months agoOpenAI promises 80% fewer hallucinations with GPT-5 debutwww.theregister.comexternal-linkmessage-square1linkfedilinkarrow-up116
arrow-up116external-linkOpenAI promises 80% fewer hallucinations with GPT-5 debutwww.theregister.comsabreW4K3@lazysoci.al to Tech@programming.devEnglish · 2 months agomessage-square1linkfedilink
minus-squarejrs100000@lemmy.worldlinkfedilinkEnglisharrow-up11·2 months agoJust from a few hours feeding it old prompts, GPT-5 seems to be about as accurate as o3 and almost as stubborn as o4. Im going to guess making it more stubborn was how they juiced the accuracy numbers.
Just from a few hours feeding it old prompts, GPT-5 seems to be about as accurate as o3 and almost as stubborn as o4. Im going to guess making it more stubborn was how they juiced the accuracy numbers.