[tomhartke]

Here's an example report on: tokenization-free large language model architectures, which have been shown to achieve compute/accuracy tradeoffs comparable to or better than traditional token-based models https://app.undermind.ai/query_app/display_one_search/05f0b8...