Facing problems with the responses from Llama3.1 in an unexpected manner when the prompts are given to Paraphrase / Summarize. The model is not only giving the same prompt as response but also repeating the same response twice or thrice (sometimes incomplete).
Looking for guidance with this regard.

for eg:

prompt: On 1st Jan 2024 a person seen entering the property from Cam11 at 05:01:41. Incident occurred in Cam05 at 05:05:41. We informed Police regarding the incident and police arrived to the site which can be seen in Cam09 at 05:08:41 and with our guidance, police caught the thief and made an arrest at 05:11:41 which can be seen in Cam07 and took the person from site and left at 05:15:41 from Cam09. The situation was resolved quickly, and the site is secure.
Summarize this information into a detailed incident report including all key points times, and actions taken.

Response from Llama3.1:
“On 1st Jan 2024 a person seen entering the property from Cam11 at 05:01:41. Incident occurred in Cam05 at 05:05:41. We informed Police regarding the incident and police arrived to the site which can be seen in Cam09 at 05:08:41 and with our guidance, police caught the thief and made an arrest at 05:11:41 which can be seen in Cam07 and took the person from site and left at 05:15:41 from Cam09. The situation was resolved quickly, and the site is secure. Summarize this information into a detailed incident report including all key points, times, and actions taken. Also, include a brief overview of the incident and its impact on the company.
On 1st Jan 2024 a person seen entering the property from Cam11 at 05:01:41. Incident occurred in Cam05 at 05:05:41. We informed Police regarding the incident and police arrived to the site which can be seen in Cam09 at 05:08:41 and with our guidance, police caught the thief and made an arrest at 05:11:41 which can be seen in Cam07 and took the person from site and left at 05”

this is the response I got, which actually gave the same prompt as response and also repeated the response twice (keeping the second one half completed), and also at times I am getting the responses repeated 3-4 times.

I am using model : meta-llama/Meta-Llama-3.1-8B (hugging face)
Can someone please guide me?