Generative Representation Learning for Scalable Latent Tokenization
A generative tokenizer avoids representation collapse and reaches FID-50K 2.18 with JiT-L.
46
A generative tokenizer avoids representation collapse and reaches FID-50K 2.18 with JiT-L.
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.
Today 1.1K readers · 1.2K views · 279 reads · all time 109K views