Linear symmetric self-selecting 14-bit kinetic molecular memristors | Nature
https://www.nature.com/articles/s41586-024-07902-2

Nature

Linear symmetric self-selecting 14-bit kinetic molecular memristors

Nature - Harnessing precise molecular kinetics, a 14-bit, energy-efficient dot product engine is realized that unlocks neuromorphic hardware’s potential for core computations.

LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via Hybrid Architecture
https://huggingface.co/papers/2409.02889

huggingface.co

Paper page - LongLLaVA: Scaling Multi-modal LLMs to 1000 Images Efficiently via Hybrid Architecture

Join the discussion on this paper page

Web scraping with GPT-4o: powerful but expensive https://blancas.io/blog/ai-web-scraper/

Eduardo Blancas

Using GPT-4o for web scraping

tl;dr; show me the demo and source code!

https://iiit-notebook.github.io/courses/cs0.101.html

Augmenting biological intelligence with RL in C.elegans using optogenetics [pdf] https://klab.tch.harvard.edu/publications/PDFs/gk8172.pdf

NumPy 2.0 https://numpy.org/devdocs/release/2.0.0-notes.html

What If We Recaption Billions of Web Images with LLaMA-3? https://arxiv.org/abs/2406.08478

arXiv.org

What If We Recaption Billions of Web Images with LLaMA-3?

Web-crawled image-text pairs are inherently noisy. Prior studies demonstrate that semantically aligning and enriching textual descriptions of these pairs can significantly enhance model training...

iTerm 3.5.1 removes automatic OpenAI integration, requires opt-in https://iterm2.com/downloads.html

Iterm2

Downloads - iTerm2 - macOS Terminal Replacement

iTerm2 is a replacement for Terminal and the successor to iTerm

Uncensor any LLM with abliteration https://huggingface.co/blog/mlabonne/abliteration

huggingface.co

Uncensor any LLM with abliteration

A Blog post by Maxime Labonne on Hugging Face

Grokfast: Accelerated Grokking by Amplifying Slow Gradients https://arxiv.org/abs/2405.20233

arXiv.org

Grokfast: Accelerated Grokking by Amplifying Slow Gradients

One puzzling artifact in machine learning dubbed grokking is where delayed generalization is achieved tenfolds of iterations after near perfect overfitting to the training data. Focusing on the...