AWhat changed
Decoupled semantic-understanding and image-generation visual encoders while retaining a shared autoregressive transformer, reducing the empirically observed representation tradeoff.
Scientific scope
Decoupled semantic-understanding and image-generation visual encoders while retaining a shared autoregressive transformer, reducing the empirically observed representation tradeoff. Inherits multimodal autoregression, discrete image tokenization and pretrained visual/LLM components. No proof that every shared encoder must fail; original Janus is separate from later flow alignment, while ordinary Pro model scaling is not another invention.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits multimodal autoregression, discrete image tokenization and pretrained visual/LLM components. No proof that every shared encoder must fail; original Janus is separate from later flow alignment, while ordinary Pro model scaling is not another invention.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction