System One Games
Copyright 2026 STEAV

The released LoRA adapters, trained decision heads, inference wrapper, and
retained synthetic decision rows are licensed under Apache License 2.0.

The encoder checkpoint and original typed decision head are from Convai Laya,
by Nandha Kishor M and contributors, licensed under Apache License 2.0:
https://huggingface.co/convaiinnovations/laya
https://github.com/NandhaKishorM/laya

The shared inference core is adapted from STEAV's published System One fraud
inference implementation (Apache-2.0); its input layout and marker scoring head
follow Laya. This repository downloads the pinned base from its upstream source.

Dino decisions were generated for an experiment using a harness derived from
the Contrastive Language Models T-Rex example. The original game engines,
planners, browser assets, and third-party harness code are not distributed here.
https://github.com/Contrastive-LM/CLM