Benchmark
#2
by
sadra
- opened
Hello, Thanks for sharing these models with MIT!
2 questions:
for 1.5 release there was a benchmark! Is the team going to share those to give a sense on improvement from 1.5 to 2 and also 2B, 4B, and 8B?
What is the source of this benchmark dataset? is it a privately held dataset or it is an open-source one?
Thanks!
Hello, yes we will release benchmark results soon. At the moment the models are not officially announced yet (we have just released them to let people to test early), that is why some information is still missing :)
Thanks! Much appreciate it!
How about the benchmark dataset itself. Will the team release information on that as well?