tulip4attoo

9 ms·

1. 1.▲
   by tulip4attoo 1y ago

   I see it normally on PC, maybe not working well on app?
2. 2.▲
   by tulip4attoo 2y ago

   I tested Carter and holy, it is so real. Sometimes I think I'm talking with a person and it's impolite to look at another screen while chatting. It's very impressive that I have to tell Carter this 2 or 3 times lol.
3. 3.▲
   by tulip4attoo 2y ago

   [flagged]
4. 4.▲
   Why you should host your LLM? From LLMOps perspective(tulip4attoo.substack.com)

   1 points by tulip4attoo 3y ago|1 comments
5. 5.▲
   by tulip4attoo 3y ago

   I don't see the reward model that they are using. I think that is an important factor when doing rlhf.
6. 6.▲
   by tulip4attoo 3y ago

   This map looks very cool. I use it on a project and it fits really well to it.
7. 7.▲
   by tulip4attoo 3y ago

   I only hit the limit when I using some bots to interact with it. Never hit it after 2 months when normally using it.
8. 8.▲
   by tulip4attoo 3y ago

   GPT4 is ways ahead. On HumanEval, it gets 67%, almost double this one.