Condensed LLM Learning with Human Feedback
By Janne Timonen, CEO
Condensed LLM Learning with Human Feedback
January 20, 2025 - Advancing AI reliability through innovative learning techniques.
Research Focus
AI is becoming increasingly essential in customer service, and this research package focuses on shaping large and small language models to be more reliable, using human feedback and clever data-condensing techniques.
Innovation Highlights
Our approach combines:
Practical Benefits
This research enables:
Methodology
By condensing training data while preserving critical information, we can train more reliable models with less computational overhead, making advanced AI more accessible.
Related posts
AI research at LastBot
LastBot's AI research focuses on language models, human feedback and continuous learning for customer communication.
Read the article
The bot said "resolved". The customer called back.
A closed conversation is not always a solved problem. Why AI-to-human handoffs need context, and how SMBs can follow up when nobody is free to take the call.
Read the article
A Bad Battery. A Good Recovery.
A comic about a five-month-old pair of headphones, a failing battery, and how connected AI customer service turned a warranty complaint into trust by checking the backend, CRM and ERP before answering.
Read the articleOffer LastBot ONE to your customers
Telecom operators, PBX vendors, software vendors and resellers can white-label, embed or resell it.