a guy created a dataset of 50 books from London 1800-1850 for LLM training. no modern bias. it’s actually super cool to see what can be trained on it!
367,7K