When to Use “Knowledge to Context” Mode

When using knowledge-enhanced chats, you typically don’t need to activate the “Enable Knowledge to Context” mode. However, there are two exceptions when you should consider enabling this feature:

1. If you’re not getting the expected results with the standard knowledge integration.
2. When the entire content of the document(s) attached to the chat is necessary to generate an accurate response.

Examples when “Enable Knowledge to Context” is not required:

1. Quick fact-checking: User: “Who wrote this report?” Attached knowledge: Contains a multi-page report. Explanation: The AI can quickly retrieve this simple fact without needing full context.

2. Summarizing a short article: User: “Summarize the main points of the article about climate change.” Attached knowledge: Contains a 500-word article on recent climate change findings. Explanation: The AI can effectively summarize the key points without needing the entire document in context.

3. Answering specific questions from a larger document: User: “What are the three main benefits of exercise mentioned in the health guide?” Attached knowledge: Contains a comprehensive health and fitness guide. Explanation: The AI can locate and extract the specific information without requiring the entire guide in context.

Examples when “Enable Knowledge to Context” is required:

1. Complex analysis of a lengthy document: User: “Analyze the themes and character development throughout the entire novel ‘To Kill a Mockingbird’.” Attached knowledge: Contains the full text of “To Kill a Mockingbird.” Explanation: The AI needs access to the entire novel to provide a comprehensive analysis.

2. Identifying patterns across multiple documents: User: “Compare and contrast the economic policies proposed in the last three State of the Union addresses.” Attached knowledge: Contains full transcripts of the last three State of the Union speeches. Explanation: The AI needs to access and analyze all three speeches in their entirety to identify patterns and make comparisons.

3. Generating a detailed report based on extensive data: User: “Create a comprehensive report on global renewable energy trends over the past decade, including regional variations and future projections.” Attached knowledge: Contains multiple lengthy reports and datasets on renewable energy from various sources. Explanation: The AI needs access to all the relevant data and reports to generate an accurate and detailed analysis.

4. Answering questions that require understanding of context and nuance: User: “Explain how the author’s personal experiences influenced the themes in their autobiography.” Attached knowledge: Contains the full text of the autobiography. Explanation: The AI needs to understand the entire narrative and context of the author’s life to draw accurate connections between experiences and themes.

In these cases where “Enable Knowledge to Context” is required, the AI needs access to the full context of the attached knowledge to provide accurate, comprehensive, and nuanced responses.