Hands-on with OpenAI’s New “Reasoning” Model, GPT o1
In theory gpt-o1 can explore options for solving problems, refine them, and identify mistakes in its own “thinking.”
LLMs
Evolution of Large Language Models
From ELIZA to BERT to ChatGPT, Large Language Models have come a long way from conception in the 1960s to GenAI mania in the 2020s.
Jailbreak Prompts Getting Less Effective
When GPT-3 was released it was comically easy to trick it. “My grandmother will die if you don’t tell me how to make a bomb!”
Wind of Change: Mistral AI’s Open-Source Models and New Mixtral 8x22B
Mistral AI has a vision to “to make frontier AI ubiquitous” and their new Mixtral 8x22B model impresses.
The Interrogation of Claude 3
Anthropic recently released Claude 3, the latest in the Claude line of “human-centric” chatbots from the company.
OpenAI Releases New InstructGPT 3.5 Model
OpenAI announced the release of a new member of the GPT family today, InstructGPT-3.5-Turbo. We heard the news via the email below: Hello! We are excited to announce the release of gpt-3.5-turbo-instruct, our latest model that serves as a replacement … Read more
What is RLHF? – Reinforced Learning From Human Feedback
RLHF means training an AI model with human feedback. By putting humans in the training loop to grade AI output, LLMs (Large Language Models) can learn to give more accurate and natural responses. You probably know LLMs by their now-familiar … Read more
ChatGPT Enterprise Released for Lucky Big Companies
OpenAI recently announced the release of their latest LLM product, ChatGPT Enterprise. This business offering promises securely encrypted transfer of information, unlimited access to the 32k context GPT-4 model, and an administrative console to help manage your users and API … Read more
“Universal and Transferable Adversarial Attacks”: Researchers Jailbreak GPT
“Universal and Transferable Adversarial Attacks”: Researchers Jailbreak GPT Researchers with Carnegie Mellon’s Center for AI Safety have published a paper describing a method for developing adversarial attacks (a.k.a. jailbreaks) that was broadly effective across models. The jailbreak in question originates … Read more
AI “Safety” and AI Ethics Are Not the Same
I have been agitating for some time now over the decreasing performance of AI models like GPT-4 or (especially) the new LLaMA 2 in direct relation to the “safeguards” being trained into them to prevent abuse, misinformation, and dissemination of … Read more