Natural stupidity vs #ArtificialIntelligence Credits: @PawlowskiMario #FridayThoughts #AI #chatgpt3 #MachineLearning #ML #chatgpt #chatbot #innovation #tech #CES2023 #FutureofWork
GENERATIVE AI
-
What is an Agent or GPT Agent Explained
By
–
What is an agent or a GPT agent? GPT agents are this emerging area where instead of going one step at a time as a prompt engineer iterating on the prompt. You let the model write X prompts and make as many commands as is required for the model to accomplish a goal you set out.
-
Logan K shares AI insights from OpenAI research
By
–
All amazing insights from @OfficialLoganK working at @OpenAI ! You can learn more in the video… https://
youtu.be/q4G6X09NEu4 And in our podcast interview: https://
open.spotify.com/episode/35SCgC
BMivXv6KeYX9yzgn?si=kK1TQwvjQHe7aLchkWjEkA
… -
What Is a Multimodal Model? GPT-4 Explained
By
–
What is a modality or a multimodal model? A multimodal model is a tongue twister. GPT4 is OpenAI's first multimodal model, which just means that it can take multiple modalities, or multiple inputs like text, and image in this case, and in the future, other modalities (eg audio).
-
Understanding Model Hallucinations in AI Systems
By
–
What is a model hallucination? Model hallucinations are when you ask a factual question, like "how many dogs are in the state of Illinois in the United States?", and the model generates something that appears to be a potentially valid answer but is entirely made up.
-
Understanding Prompts: Text Input for AI Models
By
–
What is a prompt? Prompts are just text that you send to the model.
And it can be something as simple as a question or a statement or a command, but it can also be super intricate, like a multi-paragraph or a 20 (n) step process. It's essentially what you want the model to do. -
Understanding AI Alignment: Post-Training Model Refinement
By
–
What is alignment? Alignment is a process done after training and consists of taking a base model and making it more beneficial to a human. You can go through this alignment process to make the model more in tune with what humans want to see.
-
RLHF: Aligning AI Models with Human Preferences
By
–
What is RLHF? In some sense, RLHF is part of that alignment process, where you tune the model to look at a bunch of human preferences.
You can show a bunch of example outputs and then let users, actual humans, decide what output is better for them, and you tune the model. -
Understanding Tokens: How AI Models Process Text
By
–
What is a token? Tokens are how the model understands text. Instead of understanding text at a word level or an individual character level, it understands at a token level.
-
Understanding GPT: Generative Pre-trained Transformers Architecture
By
–
What is GPT? GPT is just the architecture of the model essentially. GPT stands for Generative Pre-trained Transformers, which is the technology that GPT-4 and GPT-3.5, all those models are based on that technology.