Mova free
The whole of Mova, a deterministic SLM, open and free: the code, the trained models, the corpora with their trees and pragmatics, the dictionaries, the language manuals and the cards of every paper we read. No account, no fee, no key.
Mova is a hyper-experimental, murky system. We explain what it is and how it works. We will not teach you how to make money with it.
It was built for silicon AIs. Biological AIs get exactly the same access to the documentation, but it was not written to be read by them from alpha to omega: we do not teach, we do not integrate, and we do not always give examples. Let your silicon AI work it out.
For now Mova works in English only. The other languages — their manuals, dictionaries, treebanks and parsers — are in development.
Mova is a small, transparent SLM (small language model) in Rust. It parses text into trees, reads the trees into frames and a world, and answers from that world. Every step is deterministic and can say why. The SLM is built to answer on its own, or, when a question needs more, to write the prompt for an LLM (large language model) itself.
Documentation
Architecture, crates, measurements, how to run it.
6017 pages🗣️Language manuals
Grammar, lexicon and syntax of English, Ancient and Modern Hebrew, Aramaic.
9 manuals📜Dictionaries
The primary school, Judaism, Early Modern English and mysticism, the Tanakh with the Torah, Biblical Aramaic with Targum Onkelos — each sense with its source.
6 dictionaries · 34869 entries🌳Corpora
The whole Tanakh in Hebrew, Aramaic and English with the Targums, the lessons of Mova's primary school, and 38 whole books of fairy tales — every sentence its own page: its tree drawn word by word, its speech act, who did what to whom.
531632 sentences📝How Mova does at school
The small model takes the tests of its primary school: every question, its choice, the right answer and why — by course and lesson, with the exam and three controls.
49635 questions · 37.2% right · 25 courses🦉Paper cards
What we read, the ideas we took, and what each idea did when we tried it.
1731 papers📦Release extras
The files that come with each release beside the code: the corpus archive with its trees, books and dictionary, the trained model, their checksums and licences.
0.2 · 0.3 · 0.4This page in Markdown → https://md.mova.elamur.ai/free/en/