Patent Leaderboard
USPTO Patent Rankings Data through Dec 31, 2025
Home › Pre-training a unified natural language model with corrupted span and replaced token detection

Pre-training a unified natural language model with corrupted span and replaced token detection

US Patent 12511498 · Granted Dec 30, 2025

Assignee

Inventors

View full patent text on Google Patents →

More Patents by These Inventors

Connexin 43 (Cx43) hemichannel-binding antibodies2025
Natural language training and/or augmentation with large language models2025
Multi-least significant bit (LSB) dithering systems and methods2025
Terminal antenna and mobile terminal device2025
Electric tricycle2025
Method and application of isotropic shrinkage of three-dimensional micro-nanostr2025
Multi-domain joint semantic frame parsing2025
Signal receiving method and device, medium, and radar system2025
Fiber connector plug, fiber adapter, connector assembly, and communications devi2025
Pre-training a unified natural language model with corrupted span and replaced t2025
Pre-training a unified natural language model with corrupted span and replaced t2025
Pre-training a unified natural language model with corrupted span and replaced t2025