Welcome to basically experimental!

Here, everything is basically experimental.

This is where we publish the weird stuff. Expect alternate architectures, questionable training setups, unusual experiments, and ideas that may or may not have been a terrible idea.

What We Expect

basically experimental is for SLM research and experimentation. If you're accepted, use the organization to publish and collaborate on work related to small language models, architectures, training methods, datasets, benchmarks, or other related research.

Please don't use the organization for unrelated projects or as free file storage. Content that isn't related to SLM research may be removed, and misuse of the organization can result in your membership being terminated.

Basically: if you're experimenting with small language models, you're in the right place.