Firehose for 17 December 2024

Two repositories cover language model evaluation and agent tooling. One holds loosely organised experiments probing political censorship in large language models. The other is configuration for an assistant that redacts personally identifiable details in documents.

GitHub Repositories

2 repos created

Bias Censorship Eval Tests GitHub · Evaluation Some loosely organised experiments intended to probe the levels of political censorship inherent in LLMs
Claude Anonymisation Assistant Plugin GitHub · Agent plugin Configuration for a large language assistant designed to assist with redacting personally identifiable details in documents