Tree of thoughts
A LMQL implementation of something like tree of thoughts. Applies a natural selection process to steer reasoning and constrain the results.
Many many improvements to be made.
Features
I meant this to be as "engineerable" as possible. Each tree instance is configured to solve a specific problem, and be used as a function. It can apply a callback function to the result so it doesn't necessarily have to return text.
Some of the main features:
- Asynchronous
- Configurable
- Prompt-based and programmatic result validation
Some planned features:
- Multiple arguments and argument types
- Feature weighting: option to assign relative importance to selection criteria
- Dynamic width: method for determining how many branches should stem from each thought
How it works
Each iteration consists of a review phase, a generation phase, an evaluation phase.
- Selection: The top-k scoring lines of thought are selected
- Review: Selected lines of thought are checked to see if they contain an answer.
- Generation: A fixed number of branching thoughts are generate from selected leaf thoughts. If a selected leaf contains an answer, a conclusion is generated instead.
- Evaluation: New thoughts are scored against defined criteria to determine the relative strength of the threads. If any conclusions were generated, they are validated and returned if they pass.
Usage
For now see the examples folder to get a sense of it. In a nutshell there's three configurations: one for the initial prompt, one that governs the reasoning dynamics (evaluation, answer recognition), and one that describes how answer attempts are handled (conclusion generation, callbacks, validation).