rpglover64

PSA: Con­sider al­ter­na­tives to AUROC when re­port­ing clas­sifier met­rics for alignment

[Question] Are LLMs suffi­cient for AI take­off?

[Question] What’s ac­tu­ally go­ing on in the “mind” of the model when we fine-tune GPT-3 to In­struc­tGPT?

[Question] Image gen­er­a­tion and al­ign­ment

rp­glover64′s Shortform