Jump to content

Oracle international corporation (20250097171). LLM FINE-TUNING FOR CHATBOT

From WikiPatents

LLM FINE-TUNING FOR CHATBOT

Organization Name

oracle international corporation

Inventor(s)

Yazhe Hu of Bellevue WA US

Mengqing Guo of Issaquah WA US

Zheng Wang of Sammamish WA US

Tao Sheng of Bellevue WA US

Jun Qian of Newcastle WA US

Vinod Mamtani of Bellevue WA US

LLM FINE-TUNING FOR CHATBOT

This abstract first appeared for US patent application 20250097171 titled 'LLM FINE-TUNING FOR CHATBOT

Original Abstract Submitted

systems, methods, and other embodiments automated fine-tuning of chatbot performance for large language models are described herein. in one embodiment, a method accesses a collection of sample conversations between two entities. an individual sample conversation includes one or more rounds of natural language example prompt by a querent and example response by an agent. the method fine-tunes an llm to generate responses in natural language based on a chatbot loss function that evaluates first responses generated by the llm to the example prompts by the querent. the method generates an evaluation score for performance of the tuned llm as a chatbot based on second responses generated by the tuned llm to test prompts from a test conversation. and, the method automatically signals that the fine-tuning of the tuned llm is complete in response to the evaluation score satisfying a threshold.

(Ad) Transform your business with AI in minutes, not months

Custom AI strategy tailored to your specific industry needs
Step-by-step implementation with measurable ROI
5-minute setup that requires zero technical skills
Get your AI playbook

Trusted by 1,000+ companies worldwide

Cookies help us deliver our services. By using our services, you agree to our use of cookies.