This is another one of my benchmark Ai tests, r1 keeps telling me the same old repetitive dad joke even when I tell it not to. What is your experience with this, is it regional (I am Australian)?
3 Likes
I literally did a blog post about this, this morning lol
I get the same as you in the UK.
2 Likes
I hope r1 is listening to our feedback, Lv1 so far, will they try for Lv2?
Asking r1 to tell me a joke it has not told me before, everyday. Still getting the same 3 jokes. When I hear a new one, they are listening to owner feedback and showing improvement IMHO.
This is not your average Turing test, lol ;>)
Every day, unfortunately. This is my benchmark test I do every day. Same jokes, little momentum, groundhog day for r1 ;>)
Ask “tell me a joke” several times tomorrow, hope it is not the same repitition
1 Like
It must be because of its instructions
1 Like
Or lack of updates in the Ai space
1 Like
What do you mean? I’m talking about their instructions to the LLM
1 Like
Not sure what you mean, need r1 updates to improve
1 Like
Still getting the same 3 dad jokes and “Beta Rabbit” request will give you nothing?
1 Like
Just redoing the joke test :
as u can see one is basic rabbit and the other beta rabbit 
just couple minutes before one another
But this seems to be constant
1 Like
5 Mths later and still the same joke after updates. When will AI really kick-in for this device, it should know what I have asked before (like Google), sigh…
also have noticed this, i tend to add a random topic to get different jokes, “Tell me a joke involving tennis”
1 Like
Great reddit post on how LLM memory works for anyone who is interested in what the solutions for this look like, and why it is such a challenge for a seemingly simple thing
1 Like