title: README
emoji: 🐢
colorFrom: green
colorTo: green
sdk: static
pinned: false
BVV-LAB
Independent research in representations, memory, and modular computation.
BVV-LAB Research is an independent scientific initiative exploring how learning systems represent information, organize memory, and coordinate computation and action. Our current work focuses on language-model architectures, with a longer-term interest in robotics and embodied interaction.
Research directions
- Representations and input interfaces: distinguishing token identity from learned representations and studying fixed alternatives to trainable input tables.
- Modular architectures: investigating interfaces between contextual computation, persistent memory, and executable skills.
- Experimental methodology: evaluating capabilities, testing interfaces, and separating observed behavior from proposed explanations.
Current projects
Fixed token representations
Experiments on language modeling without independently trainable token-specific input vectors.
Modular language-model systems
Experiments with reconstructive memories, executable skills, and explicit interfaces for coordinating independently constructed components.
Open research
We publish models, code, experimental methods, and findings—including limitations and negative results—to support independent evaluation, further research, and teaching.
Each release documents its own scope, requirements, and limitations. Research prototypes should not be treated as validated deployment systems.
Existing project artifacts are currently hosted under Bochkov; this organization provides a shared home for the developing research initiative.
Collaboration
The initiative is currently maintained by Andrey Bochkov.
Contributions are welcome in independent replication, evaluation, architecture design, representation analysis, and educational materials. To discuss a specific project, please open a discussion in the relevant model repository.
When using an artifact, please consult its model card for licensing, usage conditions, and citation information.