Q36: What happens when the context sent to GitHub Copilot exceeds the LLM's maximum token limit?
When prompt context exceeds model token limits, the Copilot client employs context truncation heuristics: prioritizing active cursor lines, prefix/suffix code, open tabs, and dropping distant background files.
Key Takeaways
- Context is truncated based on proximity to active cursor.
- Active file prefix/suffix takes priority over distant repository files.
Knowledge Check
How does Copilot handle workspace context that exceeds model token capacity?