I’m not sure if you ever met in code generation use case, but sometimes in cursor (code copilot IDE), when system try to give you the code completion, it’s actually maybe the hallucinated result? e.g. a hint library that are incorrect or doesn’t ever exist
Related Topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Cursor has been hallucinating more often than usual - wasting credits | 5 | 370 | October 12, 2024 | |
GPT hallucinating | 1 | 170 | April 24, 2024 | |
Is cursor even reading the code? | 3 | 611 | September 11, 2024 | |
Seeking Insights: GitHub Codespaces with Copilot vs. Cursor for Full Codebase Context - what's the best and most reliable option as of March 2024? | 0 | 1011 | March 19, 2024 | |
Anybody else seeing hallucinations when providing document on the web? | 2 | 37 | October 13, 2024 |