OctoTools: An agentic framework with extensible tools for complex reasoning

Paper collection on building and evaluating language model agents via executable language grounding

Code & data for ICLR 2024 spotlight paper: 🍯MUSTARD: Mastering Uniform Synthesis of Theorem and Proof Data

RECKONING is a bi-level learning algorithm that improves language models' reasoning ability by folding contextual knowledge into parametric knowledge through back-propagation.

RUPBench: Benchmarking Reasoning Under Perturbations for Robustness Evaluation in Large Language Models

A curated collection of research papers on complex reasoning with language models.