Fig. 4: Scatter plots of each answer.
From: Prompt engineering in consistency and reliability with the evidence-based guideline for LLMs
a gpt-4-Web; b gpt-4-API; c gpt-4-API-0; d Bard; e gpt-3.5-Web; f gpt-3.5-API; g gpt-3.5-API-0; h gpt-3.5-ft; i gpt-3.5-ft-0.
