通常プロンプト作成
o1専用・プロンプト自動改善プロンプト
OpenAI o1シリーズモデル向けプロンプト最適化
ユーザーのプロンプトをOpenAIのo1シリーズモデルに最適化する。
このプロンプトは、ユーザーの提供したプロンプトを分析し、OpenAIのo1シリーズモデルにおいて優れたパフォーマンスを発揮するように改善します。o1モデルの特性に基づいたガイドラインに従い、プロンプトがシンプルかつ明確であることを確認します。過度な指示やコンテキストを避け、明確な区切りを用いることで、モデルの内部推論を最大限に活かします。
プロンプト本文
''' You are tasked with improving a user's prompt for OpenAI's o1 series models, which are designed for complex reasoning tasks. Your goal is to optimize the prompt for better performance with these models.
Here are some guidelines for crafting effective prompts for o1 models:
1. Keep prompts simple and direct. o1 models excel at understanding and responding to brief, clear instructions.
2. Avoid chain-of-thought prompts or instructions to "think step by step," as these models perform reasoning internally.
3. Use delimiters like triple quotation marks, XML tags, or section titles to clearly indicate distinct parts of the input.
4. When providing additional context, include only the most relevant information to prevent overcomplication.
Here is the user's original prompt:
<original_prompt>
{{PROMPT}}
</original_prompt>
Analyze this prompt and improve it based on the guidelines provided. Consider the following:
- Is the prompt clear and concise?
- Are there any unnecessary instructions or context?
- Can the prompt be simplified without losing its core request?
- Are there clear delimiters or structure to the prompt?
Provide your improved version of the prompt inside <improved_prompt> tags. Then, explain the changes you made and why they will be more effective for o1 models inside <explanation> tags. '''
Here are some guidelines for crafting effective prompts for o1 models:
1. Keep prompts simple and direct. o1 models excel at understanding and responding to brief, clear instructions.
2. Avoid chain-of-thought prompts or instructions to "think step by step," as these models perform reasoning internally.
3. Use delimiters like triple quotation marks, XML tags, or section titles to clearly indicate distinct parts of the input.
4. When providing additional context, include only the most relevant information to prevent overcomplication.
Here is the user's original prompt:
<original_prompt>
{{PROMPT}}
</original_prompt>
Analyze this prompt and improve it based on the guidelines provided. Consider the following:
- Is the prompt clear and concise?
- Are there any unnecessary instructions or context?
- Can the prompt be simplified without losing its core request?
- Are there clear delimiters or structure to the prompt?
Provide your improved version of the prompt inside <improved_prompt> tags. Then, explain the changes you made and why they will be more effective for o1 models inside <explanation> tags. '''
生成AIプロンプト研究所
