Die Epstein-Akten wurden auf dem beliebten KI-Model-Sharing-Hub Hugging Face gehostet. In der letzten Woche wurde es zum am häufigsten heruntergeladenen Datensatz auf der Plattform

Hugging Face is a platform where companies and AI researchers can host and share AI models and datasets. Last week, a user uploaded the publicly released Epstein files from the House Committee to Hugging Face, aiming to support AI powered investigative journalism.

The 20,000+ Epstein files were also made available to download a single text file on the platform [https://huggingface.co/datasets/tensonaut/EPSTEIN_FILES_20K](https://huggingface.co/datasets/tensonaut/EPSTEIN_FILES_20K)

Edit 1: Here is the Hugging face blog post on having the dataset hosted to enable open source AI tools to investigate the 20,000 documents: [https://huggingface.co/blog/tensonaut/the-epstein-files](https://huggingface.co/blog/tensonaut/the-epstein-files)

Edit 2: The data team has also started a [project space](https://github.com/EF20K/Projects) to keep track of AI apps using their dataset and to ensure safety. One of the projects that’s listed in the projects allows you to directly chat with the 20,000 documents using the open source ChatGPT model. [Here’s the link](https://agents.vectorize.io/agents/7d84aa5d-84ab-4991-b268-0b379983f57a/chat)

I always think of xenomorphs when I see that name.