[2212.13894] LAMBADA: Backward Chaining for Automated Reasoning in Natural Language - discu.eu

Reddit

[R] LAMBADA: Backward Chaining for Automated Reasoning in Natural Language - Google Research 2022 - Significantly outperforms Chain of Thought and Select Inference in terms of prediction accuracy and proof accuracy. https://arxiv.org/abs/2212.13894 11 comments 29/12/2022 machinelearning

Linking pages

Medical large language models are vulnerable to data-poisoning attacks | Nature Medicine https://www.nature.com/articles/s41591-024-03445-1 7 comments

Would you like to stay up to date with Computer science? Checkout Computer science Weekly.

Related searches:

Search whole site: site:arxiv.org

Search title: [2212.13894] LAMBADA: Backward Chaining for Automated Reasoning in Natural Language

See how to search.

Submit link to: