Personas in a Pre-Internet Language Model

David Africa & Nick LevineTLDRWe ran several persona experiments on Talkie, a 13B model trained only on English text from before 1931. We ran them alongside its modern twin (same architecture, trained on regular data as a control) and Llama-3.1-8B.Talkie can generalise maliciously. You can fine-tune Talkie on dark maxims or malicious etiquette, and it generalises to provide broadly misaligned answers.Emergent misalignment on Talkie works for some narrow datasets, such as harsh Bible verses or qu...