Metadata-Version: 2.4
Name: worldlens
Version: 0.0.1
Summary: Interpretability tools for pretrained world models
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# worldlens

Worldlens is an early-stage Python library for interpretability research on
pretrained world models. It will provide reusable interfaces and analysis tools
for inspecting how world models represent state, dynamics, goals, and other
environment factors.

The initial research code studies visual joint-embedding predictive
architectures (JEPAs), including frozen LeWM representations across manipulation
and navigation environments.

The public API is under active design and is not yet stable.
