I usually dislike memes, but given how GPT4 was before the RLHF safety work, I thought this was exceedingly appropriate and I think it really presses the point home. [image] Profoundlyyyy @profoundlyyyy OpenAI doing RLHF on the GPT-4 base model twitter.com/i/web/status/1… https://t.co/N…

GPT4 RLFH (warning: meme)

qrdl May 14, 2023, 7:58pm 3

Dark, but I couldn’t stop laughing. I have so been there.

1 Like

Topic		Replies	Views
Teknium says what we're all thinking Community chatgpt	12	655	June 25, 2024
GPT-4o Broken Security - GPT Store - Read Most Any System Prompt, Here we go again Prompting gpt-4 , chatgpt	12	2548	May 24, 2024
GPT-4 System Card by OpenAI - March 15, 2023 Community	3	5352	December 17, 2023
April 14 Livestream: Announcing GPT-4.1 series API-only models (GPT-4.5 deprecated) API developers , livestream	11	825	April 14, 2025
OpenAI did made GPT3.5 more stupid? API chatml	23	4934	December 17, 2023