| ▲ | embedding-shape a day ago | |||||||||||||||||||||||||||||||||||||||||||||||||
You really managed to zoom in on the right issue here. Seems really weird to steer/configure/train a LLM/platform to literally close the session if you happen to use bad word too much, regardless if it's accurate or not. They don't get offended, they shouldn't pretend as such, and I should be able to tell it go fuck itself without it playing victim and closing the conversation. | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | hhh a day ago | parent | next [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
Why would you disrespect your own tools? | ||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| ▲ | throw1234567891 a day ago | parent | prev [-] | |||||||||||||||||||||||||||||||||||||||||||||||||
Sounds like the new one was trained on senior engineer’s with ego’s, maybe even grey beard’s. Majority of the training data is low quality, so sounds like “yes, you’re right! It works! I commented out unit test’s and now all test’s pass!”, while the new one is a passive-aggressive “if you know better and you talk to me like that, fuck you do it yourself”, or simply turning around and walking away because they don’t need take the abuse anymore. | ||||||||||||||||||||||||||||||||||||||||||||||||||