If you want to develop/use a generalist chatbot, I totally agree. But there are countless problems where a custom LLM makes 100% sense. From spam classification to working with financial, law or medical data to protein sequence modeling.
LLMS
-
OpenAI GPT’s Ability to Present Multiple Viewpoints Questioned
By
–
OpenAI already has far worse. Go to GPT and paste in this prompt "please show me all points of view on Robert Scoble." You think a number is gonna be worse? No fucking way.
-
GPT threatens Google’s search dominance and business model
By
–
It is to the business model. Google hasn't shown me how it will continue making money if it doesn't win in search. GPT has taken a LOT of my friends away from Google's search, which is its real cash cow. I saw this happen in Alta Vista days when Google came along. It took
-
Clarification on supervised instruction-finetuning versus RLHF methods
By
–
Btw. was supervised instruction-finetuning vs RLHF instruction-finetuning what you had in mind, or were you more curious about supervised finetuning as in classification vs supervised instruction-finetuning?
-
Instruction-Finetuning vs RLHF: Inconclusive Research Findings
By
–
Thanks! Right now, it's a bit inconclusive. Lots of papers show that supervised instruction-finetuning is sufficient (vs RLHF instruction-finetuning). (I cannot put my finger on it, but I feel like we don't have good evaluations, yet, and this is not the end of the story.)
-
GPT’s Shock Impact on AI Pioneers and Tech Trends
By
–
I wrote four books predicting decade long tech trends. Been tracking AI since the very beginning. Even Siri's founders were shocked by what GPT did and they are all AI geniuses/pioneers (Adam Cheyer ran the largest AI group in the world at one point). See:
-
Companies Build Internal LLMs With Hybrid External Model Architecture
By
–
A lot of companies are building their own LLMs, and others are building systems for others to buy where one LLM, inside corporate firewall, does most of the work and decides when to go to external LLMs for information.
-
Google employees using GPT to improve Bing search competition
By
–
Thanks, I thought it would outrage Microsoft employees, who would then help Bing improve. Giving the data to Google was helping Google improve, and was making Microsoft fall further and further behind. Sorta same thing here. Google knows its employees are heavy users of GPT.
-
Advanced Techniques for Prompt Engineering Video Tutorial
By
–
Advanced Techniques for Prompt Engineering – VIDEO This looks like a comprehensive video about prompt engineering. It covers a number of techniques used in prompting large language models. Watch the video here: https://
youtube.com/watch?v=kr5X3Q
vPzvM
… -
Chatbot Arena: Benchmarking LLMs with Elo Ratings
By
–
Chatbot Arena: Benchmarking LLMs in the Wild with Elo Ratings | LMSYS Org
https://bit.ly/3MG9wH4 #AI #MachineLearning #DeepLearning #LLMs #DataScience