Task
Text Generation
This is the Dolma 3 pool, pre–quality upsampling and mixing. If you are interested in the data used to train Olmo 3 7B and Olmo 3 32B, visit allenai/dolma3mix-6T-1025.
This is the Dolma 3 pool, pre–quality upsampling and mixing. If you are interested in the data used to train Olmo 3 7B and Olmo 3 32B, visit allenai/dolma3mix-6T-1025.