The best part of ChatBotArena.

✨

WildChat-1M is now the #1 trending dataset on HF 🙀

well even if open models lose we'll still be the cool kids

This is a really cool paper comparing different pretrained LLMs. Recommend it.
Sander also helped me with a bunch of reward model stuff :)

letting my brain engage in thoughts outside the silicon valley monolith

Auto --> agents?

yo yo everyone missed this in OpenAI's model spec:
> We will also explore to what degree our models can learn directly from the Model Spec.

Constitutional AI confirmado baby🫡

Hello everyone!

Alignment Lab AI
AlignmentLab.AI,

is pleased to introduce our latest research efforts with:

Buzz!
huggingface.co/datasets/H-D-T…

a highly curated pretraining scale assistant dataset, unifying RL and SFT, developed
HIVE Digital Technologies

The Buzz model, Dataset,…