AWhat changed
Developed stable early-fusion training and mixed-modal alignment for a transformer generating interleaved discrete image/text tokens, with architectural normalization changes and a long-form mixed-modal evaluation.
Scientific scope
Developed stable early-fusion training and mixed-modal alignment for a transformer generating interleaved discrete image/text tokens, with architectural normalization changes and a long-form mixed-modal evaluation. Inherits autoregression, visual tokenization and transformer normalization. Acquired delta is the concrete stable-training/evaluation system, not first multimodal tokens or invention of QK normalization. Human preference comparisons use tested prompts/baselines and do not establish universal dominance.
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 autoregression, visual tokenization and transformer normalization. Acquired delta is the concrete stable-training/evaluation system, not first multimodal tokens or invention of QK normalization. Human preference comparisons use tested prompts/baselines and do not establish universal dominance.
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