Reproduce “Understanding the Inner Workings of a State-Space Foundation Model for Limit Order Books” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Valentin Mohl, Alexandre Bismuth, Aramis Marti-Shahandeh, George Nigmatulin, Kang Li, Stefan Zohren, Jakob Foerster, Mihai Cucuringu and 1 more
The authors test whether a message-only market model represents accumulated order flow in its recurrent state. They train an 8.2-million-parameter Mamba-3 on limit order book messages and decode order-flow imbalance (OFI) from its memory. For GOOG, the readout explains 59% of held-out 500-message OFI variation, but this is decoding, not strong future-price prediction or validated market control.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.