The BartlebyGPT Dead Letter Office (DLO) is a continued pretraining (CPT) of Qwen/Qwen3.5-2B (the Instruct SFT, not the Base model). CPT was run on ~200M tokens of Melvillian prose, over 1 epoch with Unsloth.
Downloads last month
522
Safetensors
Model size
2B params
Tensor type
BF16
·
Model tree for staeiou/bartleby-dlo-qwen3.5-2b-cpt-instructbase