AI/ML

"This collection appears to be primarily alternati…" Go see Molmo's answer!

Emails, documents, and other untrusted content can plant malicious memories.

Lessons from An Unexpected Year in AI

Large language models do better at solving problems when they show their work. Researchers are beginning to understand why.

ESM3. Enabling scientists to understand, imagine, and create proteins.

A library for graph deep learning research. Contribute to divelab/DIG development by creating an account on GitHub.

AI is just one small part of data centers’ soaring energy use.

The ultimate AI productivity app that protects your privacy. Bring all your apps and data into one AI-powered search and assistant. Get it for you and for your teams today.

Summarize and query from a lot of heterogeneous documents. Any LLM provider, any filetype, scalable, under developpement - thiswillbeyourgithub/WDoc

I decided to have a poke around and see if I could figure out how the HTTP streaming APIs from the various hosted LLM providers actually worked. Here are my notes so far.

Finetuning a GPT Model for Spam Classification

Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

LLaMA-Omni is a low-latency and high-quality end-to-end speech interaction model built upon Llama-3.1-8B-Instruct, aiming to achieve speech capabilities at the GPT-4o level. - ictnlp/LLaMA-Omni

Sourcetable is an AI spreadsheet that helps you analyze data and create reports. Chat with your data, create charts and graphs, build financial models, + more.

Empowering RAG with a memory-based data interface for all-purpose applications! - qhjqhj00/MemoRAG

Here's an interesting new embedding/RAG technique, described by Anthropic but it should work for any embedding model against any other LLM. One of the big challenges in implementing semantic search …

Introduction Every single time I design a protein — using ML or otherwise — I am confident that it is capable of being manufactured. I simply reach o…

What productionizing test-time compute shows us about the future of AI. Exploration has landed in language model training.

Things you can do with the token embeddings of an LLM - dleemiller/WordLlama

Reader-LM-0.5B and Reader-LM-1.5B are two novel small language models inspired by Jina Reader, designed to convert raw, noisy HTML from the open web into clean markdown.

Co-author Gordon Pennycook: "The work overturns a lot of how we thought about conspiracies."

OpenAI released two major new preview models today: o1-preview and o1-mini (that mini one is also a preview, despite the name)—previously rumored as having the codename “strawberry”. There’s a lot …

New version of my `files-to-prompt` CLI tool for turning a bunch of files into a prompt suitable for piping to an LLM, [described here previously](https://simonwillison.net/2024/Apr/8/files-to-prompt/). It now has a `-c/--cxml` …

Yes, the rumours are true! Kagi has been thoughtfully integrating AI into our search experience, creating a smarter, faster, and more intuitive search.

Build Spirals with your voice, personality, and style to instantly start at 80% done on any repeat writing, thinking, or decision-making task.

What comes after the Homework Apocalypse

New hosted API for Llama running at absurdly high speeds: "1,800 tokens per second for Llama3.1 8B and 450 tokens per second for Llama3.1 70B". How are they running so …

Anthropic has published the system prompts for Claude, its family of generative AI models. They instruct the model how it should -- and shouldn't -- behave.

Anthropic now publish the system prompts for their user-facing chat-based LLM systems - Claude 3 Haiku, Claude 3 Opus and Claude 3.5 Sonnet - as part of their documentation, with …