> It’s not dead. Some companies use it while I believe the majority of use cases are addressed by open-weight rivals.
The quality of the answers on my datasets are in another league compared to the benchmarks.
I have a mixed language dataset, Jev answers only lack some accuracy, the open weight rivals output just look like random noise
The quality of the answers on my datasets are in another league compared to the benchmarks. I have a mixed language dataset, Jev answers only lack some accuracy, the open weight rivals output just look like random noise