You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm using the llama-3 70b model for an application. However, I've noticed that the model tends to repeat the same phrases or response structures frequently, even when given slightly different inputs. This is impacting the user experience, as it makes the dialogue predictable and unnatural.
Example:
when i introduce myself, it keeps adding "!" at the beggining of the answers
Has anyone else encountered this issue? Any suggestions on how to improve the diversity in responses?
The text was updated successfully, but these errors were encountered:
I'm using the llama-3 70b model for an application. However, I've noticed that the model tends to repeat the same phrases or response structures frequently, even when given slightly different inputs. This is impacting the user experience, as it makes the dialogue predictable and unnatural.
Example:
when i introduce myself, it keeps adding "!" at the beggining of the answers
Has anyone else encountered this issue? Any suggestions on how to improve the diversity in responses?
The text was updated successfully, but these errors were encountered: