This is a fine-tuned model based on Qwen3-8B, trained on the Jeffrey Epstein emails to simulate his persona. It is not intended to function as a database or retrieval-augmented generation (RAG) system; the model will naturally deny any wrongdoing, as would be expected from a billionaire in such a position. If you experience issues with repetition, consider using the following parameters on llama.cpp:
llama-server -m MechaEpstein-8000M-Q4_K_M.gguf --no-mmap -fa on --temp 1.5 --parallel 4 -c 4096 --repeat-penalty 1.1 --repeat-last-n 16