Cobus GreylingData Design For Fine-Tuning LLM Long Context WindowsFine-Tune your LLM To fully utilise the available context window.5 min read·2 days ago----
Cobus GreylingThe Importance Of Granular Data Design For Fine-TuningAnd leveraging Data Design to train LLMs to fully utilise context, while also solving for the Lost-In-The-Middle challenge.6 min read·3 days ago----
Cobus GreylingLangChain Structured Output Parser Using OpenAIThe LangChain structured output parser in its simplest form closely resembles OpenAI’s function calling. In this article I include a…5 min read·5 days ago--1--1
Cobus GreylingThree Considerations For Private Open-Source LLM InstancesUsing commercial LLM APIs for production applications pose clear & well researched risks. Consequently, enterprises are increasingly…6 min read·6 days ago--1--1
Cobus GreylingIntents Are Not Going Away…RoNID Is A New Intent Discovery FrameworkThe Robust New Intent Discovery (RoNID) framework strives to identify known intents and reasonably deduce new intent groups in open domain…5 min read·Apr 26, 2024----
Cobus GreylingThe Large Language Model Landscape — Version 5In the recent past I have been observing and describing current LLM-related technologies and trends. In this article I’m taking a step…5 min read·Apr 25, 2024--2--2
Cobus GreylingLLMs Excel At In-Context Learning (ICL), But What About Long In-context Learning?This study was inspired by the Lost in the Middle phenomenon which surfaced in 2023.5 min read·Apr 24, 2024----
Cobus GreylingUsing LLMs For Autonomous VehiclesLC-LLM is an explainable lane change prediction model that leverages the strong reasoning capabilities & self-explanation abilities of…6 min read·Apr 23, 2024----
Cobus GreylingMatching Retrieved Context With Question Context Using LogProbs With OpenAI for RAGTo minimise hallucinations and enhance the efficiency of RAG-powered Question Answering systems, with OpenAI log probabilities can be used…8 min read·Apr 22, 2024----
Cobus GreylingRapid Development Of Intelligent Generative AI APIsThere are instances where merely exposing an API to an LLM does not suffice, and some intelligence needs to be added to the API. Here is a…4 min read·Apr 19, 2024----