Least-to-Most Prompting Enables Complex Reasoning in Large Language Models

Open paper
Benchmark Model Rank Results
arithmetic-reasoning-on-gsm8kcode-davinci-002 (Least-to-Most Prompting)#75Accuracy: 68.01Parameters (Billion): 175
question-answering-on-strategyqaLeast-to-Most#7EM: 65.8

Rank counts only results with a code link.