Turn any LLM into a Jev-style decision model: typed decisions, real probabilities, no training.
Why it matters now: Turn any LLM into a Jev-style decision model
AnyJev is a tool that allows users to turn any Large Language Model (LLM) into a Jev-style decision model. This means that the model can produce typed decisions, real probabilities, and does not require any fine-tuning. The tool is designed to be easy to use and can be integrated into existing workflows. AnyJev supports various levels of decision-making, including L2, which reads a hidden state and can be served using a pooling server plus a few kilobytes of head. The tool also provides a simple way to convert, serve, and measure models in one command using the `python -m anyjev.pipeline` command.
Python and a compatible LLM model