From prompt attacks to data leaks, LLMs offer new capabilities and new threats
The Stack Overflow Podcast
00:00
Experiments with Fine-tuning LLMs and Source Attribution
The speakers share their experiences with fine-tuning large language models on specific datasets, including the use of 'llama v2' and a model connected to online content. They achieved improved results and conducted the experiments on their laptops to avoid expensive cloud services.
Transcript
Play full episode
