Salesforce, Inc. patent applications on April 3rd, 2025
Patent Applications by Salesforce, Inc. on April 3rd, 2025
Salesforce, Inc.: 2 patent applications
Salesforce, Inc. has applied for patents in the areas of G06F40/284 (2), G06N3/0455 (1), G06F40/40 (1), G06N3/047 (1) G06F40/284 (1), G06N3/0455 (1)
With keywords such as: output, text, probability, neural, input, network, next, token, based, and generated in patent application abstracts.
Patent Applications by Salesforce, Inc.
Inventor(s): Tong Niu of Sunnyvale CA US for salesforce, inc., Yingbo Zhou of Palo Alto CA US for salesforce, inc., Silvio Savarese of Palo Alto CA US for salesforce, inc., Semih Yavuz of Redwood City CA US for salesforce, inc., Caiming Xiong of Menlo Park CA US for salesforce, inc.
IPC Code(s): G06F40/284
CPC Code(s): G06F40/284
Abstract: embodiments described herein provide a method for mitigating toxic content in text generation by a neural network based framework. the method includes the following operations. a text input of a sequence of tokens is received via a communication interface. a first output probability for a next token generating is generated by a first neural network model that is trained to generate tokens belonging to a prioritized category of vocabulary, in response to the text input. a second output probability of the next token is generated by a second neural network model that is trained to generate tokens belonging to an indiscriminate vocabulary, in response to the text input. the next token for a text output based on a combined output probability computed based on a correction item reflective of the first output probability and the second output probability is generated in response to the text input.
Inventor(s): Lifu TU of San Francisco CA US for salesforce, inc., Semih YAVUZ of Redwood City CA US for salesforce, inc., Yingbo ZHOU of Palo Alto CA US for salesforce, inc.
IPC Code(s): G06N3/0455, G06F40/284, G06F40/40, G06N3/047
CPC Code(s): G06N3/0455
Abstract: in view of the need to improve text generation technology, embodiments described herein provide a neural network model that generates a text output with constraints to achieve desired output behavior, such as reduced toxicity or hallucinations, and inclusion of certain keywords.