Un outil en ligne (qui tourne entièrement dans le navigateur) pour extraire le texte d'images ou de PDF. (simonwillison.net) French
The GPT-4 barrier has finally been broken (simonwillison.net)
Four weeks ago, GPT-4 remained the undisputed champion: consistently at the top of every key benchmark, but more importantly the clear winner in terms of “vibes”. Almost everyone investing serious time exploring LLMs agreed that it was the most capable default model for the majority of tasks—and had been for more than a...
The AI trust crisis (simonwillison.net)
llamafile is the new best way to run a LLM on your own computer (simonwillison.net)
Deciphering clues in a news article to understand how it was reported (simonwillison.net)
Making Large Language Models work for you (simonwillison.net)
Git scraping: track changes over time by scraping to a Git repository (simonwillison.net)
Simon Willison’s LLM CLI tool now supports self-hosted language models via plugins (simonwillison.net)
LLM is my command-line utility and Python library for working with large language models such as GPT-4. I just released version 0.5 with a huge new feature: you can now install plugins that add support for additional models to the tool, including models that can run on your own hardware....
My LLM CLI tool now supports self-hosted language models via plugins (simonwillison.net)
Understanding GPT tokenizers (simonwillison.net)
This is an excellent overview of tokenization with many interesting examples. I also like Simon’s small CLI tools; you can read about them at the end of the post....
symbex: search Python code for functions and classes, then pipe them into a LLM (simonwillison.net)
From the article:...
The Dual LLM pattern for building AI assistants that can resist prompt injection (simonwillison.net)
An interesting and clever proposal to fix the prompt injection vulnerability....
Understanding GPT tokenizers (simonwillison.net)
Saw this on HN and thought it was very interesting. Also wanted to test creating a post on Lemmy :)
Leaked Google document: “We Have No Moat, And Neither Does OpenAI” (simonwillison.net)
Interesting article about how open source LLMs are picking up pace compared to proprietary ones.
It’s infuriatingly hard to understand how closed models train on their input (simonwillison.net)
Leaked Google document: “We Have No Moat, And Neither Does OpenAI” (simonwillison.net)
"...The premise of the paper is that while OpenAI and Google continue to race to build the most powerful language models, their efforts are rapidly being eclipsed by the work happening in the open source community..."
symbex: search Python code for functions and classes, then pipe them into a LLM (simonwillison.net)
cross-posted from: https://programming.dev/post/107386...
Delimiters won’t save you from prompt injection (simonwillison.net)
Prompt injection remains an unsolved problem. The best we can do at the moment, disappointingly, is to raise awareness of the issue. As I pointed out last week, “if you don’t understand it, you are doomed to implement it.”