Google BARD is Out!
Google BARD just rolled out their own experimental site for their ChatGPT and GPT-4 rival, BARD.
Created on March 22|Last edited on March 22
Comment
Google BARD just rolled out their own experimental site for their ChatGPT and GPT-4 rival, BARD. You can join the waitlist here.
Sparrow, Claude, LamDA, PaLM, LLaMA, Alpaca, Bard, ChatGPT, and GPT-4. LLMs are taking over the deep learning and AI space. With so many rivals, it's natural to start comparing the performances of these models. I recently saw this video:
This video showed a few shortcomings of BARD where it wasn't able to correctly answer or summarize certain questions, topics GPT-4 succeeded in. Though do keep in mind this set of experiments was just an initial peek into BARD.
It's interesting to think that as these LLMs reach new heights, our method of measuring their supposed intelligence grows vaguer and vaguer. The GPT-4 Report evaluated the model on AP Exams and standardized tests! It's difficult to evaluate a model's intelligence when it generates outputs so well. There's another aspect to it: how do we measure intelligence, critical thinking, and reasoning? I think in the coming months we might see a new set of practices for monitoring these models in production and evaluating their performances such that it aligns with human virtue.
References
Mihalcik, Carrie, and Imad Khan. “Google Says You Can Now Try out Bard, Its Chatgpt Rival.” CNET, CNET, 22 Mar. 2023.
Add a comment
Tags: ML News
Iterate on AI agents and models faster. Try Weights & Biases today.