⛓️ Answering Questions by Meta-Reasoning over Multiple Chains of Thought

Multi-Chain Reasoning (MCR) allows answering complex by introducing a meta-reasoner which reasons over multiple CoTs. MCR can mainly be useful in answering complex questions in which multiple answering strategies are possible.

MCR System Overview

🤖 Running MCR

If you are interested in running MCR, check out our demo notebook.

⛰️ Datasets and evaluation

To evaluate MCR, we provide our dataset readers and evaluators. Code to run end-to-end experiments be released soon.

✍ Citation

bibtex
@article{yoran-etal-2023-answering,
    title={Answering Questions by Meta-Reasoning over Multiple Chains of Thought}, 
    author={Ori Yoran and Tomer Wolfson and Ben Bogin and Uri Katz and Daniel Deutch and Jonathan Berant},
    year={2023},
    eprint={2304.13007},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}