Studied by 0 people
Looks like no one added any tags here yet for you.
Generative AI
Learns about the patterns in training data and generates new content based on those patters after they receive input like a text prompt
can create new content like images, audio, and code
Large Language Models (LLM)
Type of generative AI that is pre-trained w/ a very large set of data and then gets fine-tuned for specific tasks
ChatGPT (Generative Pre-Trained Transformer)
Type of LLM that can create human-like text by understanding sentence structure and meaning
was released by OpenAI in 2022
Predicts responses based on learned patterns rather than retrieving answers from the internet
Emergent Properties
AI sometimes develops unexpected skills it wasn’t explicitly trained for
Ex: creating a human tale w/ characters it invented from a 6-word prompt
GPT 4o
ChatGPT for high interaction rates and better language understanding + ability to produce more detailed responses
Cannot “learn on the fly” → must be taken offline for updates
Can search the internet for real-time information
uses voice chat integration, writes meeting summaries, analyze spreadsheets, code generation, interview prep
Hallucinate
Generation of a response that is incorrect or fabricated but sounds plausible
AI models often overconfidently justify incorrect answers
Reasoning Models
Types of models that generate different possible answers or options privately, then evaluates the viability of each of those before responding to the user w/ what it deems to be the best answer
