A hobbyist AI developer experimenting with Victorian-era language models “just for fun” received a surprising history lesson when his latest model generated text referencing actual London protests from 1834—events he didn’t know existed until he looked them up.
Reddit user Hayk Grigorian, a computer science student at Muhlenberg College in Pennsylvania, shared the discovery online after his model unexpectedly cited real historical unrest. “I was interested to see if a protest had actually occurred in 1834 London and it really did happen,” he wrote.
For about a month, Grigorian has been building TimeCapsuleLLM, a small custom language model trained exclusively on texts from London between 1800 and 1875. His goal is to recreate an authentic Victorian voice, resulting in outputs rich with biblical allusions, elaborate rhetoric, and period-specific writing styles.
TimeCapsuleLLM is part of a growing research trend exploring “Historical Large Language Models,” which train AI on writings from specific eras to simulate historical perspectives. Comparable projects include MonadGPT, built on 11,000 texts from 1400–1700 CE and capable of reasoning with 17th-century conceptual frameworks, and XunziALLM, which produces classical Chinese poetry following ancient stylistic rules.
The most surprising output came from a simple test prompt: “It was the year of our Lord 1834.” Trained to continue text in its learned style, the model generated the following passage:
It was the year of our Lord 1834 and the streets of London were filled with protest and petition. The cause, as many re counted, was not bound in the way of private, but having taken up the same day in the day of Lord Palmerston, the public will receive a short statement of the difficulties under which the day of law has reached us. It is a matter of deep regret, that the present events in the history of the world are clear, and consequently will be ’known. It is not true that the very men who first settled in the Gospel at Jerusalem should have so extensive and so interesting a record of the prosperity and prosperity
Intrigued, Grigorian began fact-checking the generated details. Not only had the model referenced real mid-1830s political unrest, but it also correctly mentioned Lord Palmerston, whose policies were indeed linked to protests in 1834. “The output also brought up Lord Palmerston,” he wrote, “and after a Google search I learned that his actions resulted in the 1834 protests.”