Introducing Contextual Retrieval \ Anthropic
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
AI/ML
GitHub - ictnlp/LLaMA-Omni: 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.
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 | Your AI Data Analyst
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.
qhjqhj00/MemoRAG: Empowering RAG with a memory-based data interface for all-purpose applications!
Empowering RAG with a memory-based data interface for all-purpose applications! - qhjqhj00/MemoRAG
Introducing Contextual Retrieval
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 …
Generative ML in chemistry is bottlenecked by synthesis — LessWrong
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…
Reverse engineering OpenAI’s o1
What productionizing test-time compute shows us about the future of AI. Exploration has landed in language model training.
dleemiller/WordLlama: Things you can do with the token embeddings of an LLM
Things you can do with the token embeddings of an LLM - dleemiller/WordLlama
Reader-LM: Small Language Models for Cleaning and Converting HTML to Markdown
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.
AI chatbots might be better at swaying conspiracy theorists than humans
Co-author Gordon Pennycook: "The work overturns a lot of how we thought about conspiracies."
Notes on OpenAI’s new o1 chain-of-thought models
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 …
files-to-prompt 0.3
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` …
Announcing The Assistant | Kagi Blog
Yes, the rumours are true! Kagi has been thoughtfully integrating AI into our search experience, creating a smarter, faster, and more intuitive search.
Start at 80% done on any writing, thinking, or creative task.
Build Spirals with your voice, personality, and style to instantly start at 80% done on any repeat writing, thinking, or decision-making task.
Post-apocalyptic education
What comes after the Homework Apocalypse
dh1011/llm-term
A Rust-based CLI tool that generates and executes terminal commands using OpenAI's language models or local Ollama models.
Cerebras Inference: AI at Instant Speed
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 publishes the 'system prompt' that makes Claude tick | TechCrunch
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 Release Notes: System Prompts
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 …
Long context prompting tips
Interesting tips here from Anthropic's documentation about how to best prompt Claude to work with longer documents. **Put longform data at the top**: Place your long documents and inputs …
Episode 6 of Shell Game: The Future Isn't Real
Shell Game Episode 6, our final episode of our first season, is out now!
While at times it may have seemed like I was just using my voice agent to mess around and have some fun—in fairness, many times I was—you’ve hopefully detected by now that we were also on an intentional journey over the course of the first five episodes. From a voice clone that knew nothing about me in its calls to customer service, to one built to tie up scammers. Then from an agent that could represent me in my most superficial conversation, to one that could express my inner life, and try and take on my job.
In the last episode, we arrive at our destination, where a cheshire cat awaits us, with a pair of questions. First, a general one: How do we grasp the benefits of new technologies, like AI voice, without simultaneously letting go of things that matter to us? And as a way of approaching that question, another: What does it feel like for the people closest to me to confront AI me, instead of me?
To answer them, I had to push this experiment as far as I was comfortable and then beyond. This was the hardest episode to make, for reasons that will become clear when you hear it. But I hope that you feel the way the three of us working on it have felt, that it will stick with you for a long time.
An Age of Hyperabundance | Laura Preston
Everyone at this conference kept invoking loneliness and claiming the antidote was conversation. That didn’t track with my own experience. My most desperate moments of loneliness have been in conversation: on a Hinge date, doomed but persisting as a form of protocol. At a publishing party, surrounded by people who look and talk like me, all of us a little drunk but maintaining our nervous, manic professionalism.
Postgres as a search engine
Build a retrieval system with semantic, full-text, and fuzzy search in Postgres to be used as a backbone in RAG pipelines.
Create Calendar Entries with Anthropic Claude 3.5
Generate calendar entries from an image using Anthropic Claude 3.5. Convert lesson dates in an image to a calendar ICS file
Exploring recipes with LLMs, Ollama and R – Thomas Sandmann’s blog
Evaluating the Effectiveness of LLM-Evaluators (aka LLM-as-Judge)
Use cases, techniques, alignment, finetuning, and critiques against LLM-evaluators.
What’s Really Going On in Machine Learning? Some Minimal Models
Stephen Wolfram explores minimal models and their visualizations, aiming to explain the underneath functionality of neural nets and ultimately machine learning.
%2Fcdn.vox-cdn.com%2Fuploads%2Fchorus_asset%2Ffile%2F25584373%2F247226_AI_Pixel_Reality_CVirginia_3.jpg?width=92&height=&ar=&mode=&format=avif&dpr=2){kind=tracking}
No one’s ready for this
The photograph is now meaningless as evidence. We are not prepared.
Fine-tuning now available for GPT-4o
Fine-tune custom versions of GPT-4o to increase performance and accuracy for your applications
light-the-torch
`light-the-torch` is a small utility that wraps `pip` to ease the installation process for PyTorch distributions like `torch`, `torchvision`, `torchaudio`, and so on as well as third-party packages that …