Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
Vector<Float> -> Tensor<Embedding>
Map continuous robot state estimation vectors into LLM input embeddings using a multi-layer perceptron.
Problem it solves
Raw continuous state estimations cannot be natively tokenized by discrete vocabulary models.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.