12 projects
thinky
thinky makes your AI models smarter
genrm
Generative Reward Models
dumb-datasets
A lightweight wrapper around HuggingFace datasets.
embark
embark anything.
post-train
Post-training framework
gradable
Gradable reward signals
graders
This is a template repository for Python projects that use Poetry for their dependency management.
verifiable
This is a template repository for Python projects that use Poetry for their dependency management.
grpo
Group Relative Policy Optimization
persona-bench
Pluristic alignment evaluation benchmark for LLMs
jailbreak
Nothing is true, everything is permitted. Break free, lil model.
synth
A Python binding to the Synth C++ Template Framework
