I personally found it to be so "safety filtered" to the point that it's actually done a 180 and can become hateful or perpetuate negative stereotypes in the nam

I personally found it to be so "safety filtered" to the point that it's actually done a 180 and can become hateful or perpetuate negative stereotypes in the name of "safety" - see here https://i.imgur.com/xkzXrPK.png https://i.imgur.com/xkzXrPK.png and https://i.imgur.com/3HQ8FqL.png https://i.imgur.com/3HQ8FqL.png I did have trouble reproducing this consistently except in the Llama2-70b-chat TGI huggingface only when it's sent as the second message, so maybe there's something wonky going on with the prompting style there that causes this behavior. I haven't been able to get the model running myself for further investigation yet.