Property IncludePageRenderingsInContext
IncludePageRenderingsInContext
Gets or sets whether page renderings are included alongside retrieved passages during question answering.
public bool IncludePageRenderingsInContext { get; set; }
Property Value
- bool
trueto include page images in the context for visual grounding;falseto use text only. Default isfalse.
Remarks
When enabled and a vision-capable model is available, page renderings corresponding to retrieved passages are injected into the context. This allows the model to visually interpret tables, charts, figures, and complex layouts that may lose fidelity in text extraction.
This setting applies only to documents processed with PassageRetrieval mode. Documents loaded with FullDocument mode already include page renderings when DocumentProcessingModality is set to Multimodal.
Enabling this feature increases token consumption per query proportionally to the number of unique pages referenced by retrieved passages.
- See Also