Talk to CaLLM Why Climate Change Matters What, Why and Who is CaLLM? Large Language Models (LLMs) are a type of artificial intelligence that is able to generate human-like text on a wide range of topics. The progress that has been seen in recent years in the realm of LLMs is staggering. The development and release of models such as GPT- 3, Llama2, PaLM2 and GPT-4 have already begun to completely change the natural language processing landscape. This allows for huge leaps forward in a plethora of fields, from research to marketing, from email writing to education. However, there are also a range of limitations that exist at the forefront of development. LLMs are prone to ‘hallucinations’ that is, they produce believable, human-like text that is completely false in content. This is a very substantial drawback in contexts where the topics being discussed are sensitive and/or controversial. The problem of ‘fake news’ is rampant enough in real human discourse; we should limit it where we can in artificial human discourse. There are also many ethical and philosophical deliberations to be considered surrounding bias in training data, a lack of transparency in development and the idea that a machine is creating human-like output. CaLLM is an attempt to address some of these issues. He is a guide that is powered by the open-source Meta AI Llama2 large language model. This means he has been trained on their large dataset. However, to increase his specificity and knowledge base, I have used a Retrieval Augmentation Generation technique (RAG) so that we can add to his repository of known information that he is able to call on when answering a question. This is all to try and provide some accuracy and precision in the use of LLMs, particularly around essential conversations including climate change and heritage. For more information on how this was achieved and more information on the RAG technique, check out my GitHub and to learn more about Llama2, take a look at their site. GitHub Llama2