New blog post up, and it's the hottest take I've ever made. I've been using the ChatGPT API for the past few days, and it's the real deal to the point that I can't justify using anything else for AI text generation. minimaxir.com/2023/03/new-ch…

This is a 💯 idea. I wonder if it's possible to more generally filter out "low perplexity" text using a less powerful LM

Just added support for all LLaMA models I'm out of disk space, so if someone can give this a try for 33B and 65BB would be great 😄 See updated instructions in the Readme Here is LLaMA-13B at ~10 tokens/s

Instruction-following models are now ubiquitous, but API-only access limits research. Today, we’re releasing info on Alpaca (solely for research use), a small but capable 7B model based on LLaMA that often behaves like OpenAI’s text-davinci-003. Demo: crfm.stanford.edu/alpaca/

I’ve been building stuff w agents for ~3 months. Some scattered thoughts on the current agent hype: On agent reliability: before you get too scared or hyped about autonomous agents, try running one yourself they have flashes of clarity and then tend to rabbit hole / get stuck

3. Arcwise AI: A plugin for Google Sheets that instantly understands, cleans, and ingests data in your sheets using AI. Ask Arcwise any question about your spreadsheet and it will do the math for you and give you the right answer in seconds:

Introducing: 💫StarCoder StarCoder is a 15B LLM for code with 8k context and trained only on permissive data in 80+ programming languages. It can be prompted to reach 40% pass@1 on HumanEval and act as a Tech Assistant. Try it here: shorturl.at/cYZ06r Release thread🧵

Will future LLMs be based almost entirely on synthetic training data? In a new paper, we introduce TinyStories, a dataset of short stories generated by GPT-3.5&4. We use it to train tiny LMs (< 10M params) that produce fluent stories and exhibit reasoning. arxiv.org/abs/2305.07759

QLoRA: 4-bit finetuning of LLMs is here! With it comes Guanaco, a chatbot on a single GPU, achieving 99% ChatGPT performance on the Vicuna benchmark: Paper: arxiv.org/abs/2305.14314 Code+Demo: github.com/artidoro/qlora Samples: colab.research.google.com/drive/1kK6xasH… Colab: colab.research.google.com/drive/17XEqL1J…

this is super fascinating, it’s almost as if the internal representation behaves like a hash table 🤔

i've been dreaming of an AI generator for Mantine components for a while! Patterns nailed my first request for a pretty boilerplate modal 👏