These papers use some sort of reasoning chain, including natural language reasons or natural language “rules.”

• EntailmentBank: Dalvi et al 2021 - Explaining Answers with Entailment Trees Open domain, “1,840 multistep entailment trees for accompanying QA pairs”
• Tafjord et al 2021 - ProofWriter: Generating Implications, Proofs, and Abductive Statements over Natural Language
• Dalvi et al 2022 - Towards Teachable Reasoning Systems This paper v1 got split up into two papers (both at EMNLP 2022):
  • Tafjord et al 2022 - Entailer: Answering Questions with Faithful and Truthful Chains of Reasoning
  • Dalvi et al 2022 - Towards Teachable Reasoning Systems: Using a Dynamic Memory of User Feedback for Continual System Improvement
• Sprague 2022 - Natural Language Deduction with Incomplete Information

See Commonsense Reasoning - Abductive Reasoning.

• EntailmentBank: paper Open domain, “1,840 multistep entailment trees for accompanying QA pairs”

• Peter Clark

• NLI
• Question Answering