Hello, the parameter setting is asia-250-1-softHC-bic, and the model uses gpt-3.5. However, the final result is always the best when there is no LLM constraint, that is, when LLM is not used. Is it possible that there is an issue with how I'm running the code? Additionally, even in the best result, there is still an incorrect edge. I noticed in your paper that the accuracy for asia-250 is reported as 100%. I would appreciate your guidance on this matter. Thank you very much!
Hello, the parameter setting is asia-250-1-softHC-bic, and the model uses gpt-3.5. However, the final result is always the best when there is no LLM constraint, that is, when LLM is not used. Is it possible that there is an issue with how I'm running the code? Additionally, even in the best result, there is still an incorrect edge. I noticed in your paper that the accuracy for asia-250 is reported as 100%. I would appreciate your guidance on this matter. Thank you very much!