enpo-llama3-8b-eta-eighth-seed1-cuda13-iter1-adversary

An experimental fine-tuned or adversarially-configured checkpoint derived from Meta's Llama 3 8B base model, uploaded by the ntlfi account on Hugging Face. The naming convention — including fields such as eta-eighth, seed1, cuda13, iter1, and adversary — strongly indicates an internal research or evaluation run rather than a productized or publicly released model. Source: supabase.