chunked/content_aware_chunking/_model_summary/chunk_19.txt

GPT-J was followed by OPT, a family of decoder-only models, the largest of which is 175B and trained on 180B tokens.