Excited to share our new dataset, FOL-Traces!
We introduce a large-scale dataset of programmatically verified FOL reasoning traces for studying structured logical inference + process fidelity.
Happy to hear thoughts from others working on reasoning in LLMs!
Check it out here 👇