Ethan Mollick highlights persona experiments with a pre-1930 model

What happened

There have been a bunch of really interesting experiments with activating individual neurons in open AI models, locking them into particular personas. Here historian @ResObscura tries the same trick with a model trained only on data before 1930, and the results are delightful. https://twitter.com/resobscura/status/2115176788993380661

Read next

• More changes in today’s briefingGoogle DeepMind announces Genie 4 for real-time world simulation with native audio

The story so far

Ethan Mollick highlighted historian @ResObscura’s experiments activating individual neurons in Talkie-1930 to induce particular personas, describing the results as “delightful.” Mollick said the model was trained only on data from before 1930 and connected the experiment to similar work using neuron activation to steer personas in open AI models.