Hidden State Research is an independent research initiative working on machine learning and intelligent systems.
Our goal is to explore new ideas, turn them into practical systems, and share what we learn through publications and open-source projects.
NeurIPS 2026 Workshop on Agentic AI Benchmarks and Applications for Enterprise Tasks (AABA4ET)
A benchmark for evaluating how reliably open-weight agents carry out internal enterprise operations tasks.
Visualize PyTorch neural network architectures in a few lines of code.
Generate layered, graph and LeNet-style architecture diagrams straight from a torch.nn.Module. Useful for papers, slides and documentation.