LLMS

→ View original post on X — @aibreakfast

Google PaLM/Minerva Outperforms Students on Polish Math Exam

By

@aibreakfast

–

29 December 2022 4h51

Google PaLM/Minerva outperformed students on the 2020 Polish math exam. A real question from the test, and subsequent correct answer by AI.

→ View original post on X — @aibreakfast

GPT-based Plagiarism Detection Tool on Huggingface

By

@aibreakfast

–

29 December 2022 4h34

GPT “plagiarism” detector https://
huggingface.co/openai-detecto
r/
…

Automating Prompts for User-Friendly AI Applications

By

–

29 December 2022 2h30

Features like this should be ubiquitous. Low-risk applications where you need to present some label to the user — naming a file, a conversation, a revision of a file. All these things can be described succinctly and automated in a prompt. Our apps should speak our language.

Integrating LLMs into Applications and the Potential of Scale Spellbook

By

–

29 December 2022 2h25

For all the attention ChatGPT and chat interfaces get, we haven't digested what's possible with this year's LLMs. We haven't finished integrating LLMs into our apps. Getting there is a big part of what I want to achieve through Scale Spellbook. https://
scale.com/spellbook

Copilot’s Grounding in Codebase via Prompt Engineering Explained

By

–

29 December 2022 0h23

Copilot also has lots of grounding in the codebase through clever prompt engineering so that might be what you’re noticing.

ChatGPT Hallucinations Reduced Compared to GPT-3

By

–

29 December 2022 0h15

ChatGPT hallucinates a lot less. It makes stuff up in the details all the time, but normal GPT-3 will define nonsense words or imagine the text of fake URLs. See here for examples: