| base_model: unsloth/Qwen3-14B-unsloth-bnb-4bit | |
| library_name: peft | |
| tags: [security, self-improvement, self-play, alien-si, rsi] | |
| license: apache-2.0 | |
| # ALIEN SI SELF-PLAY — 14B (Frontier Level) | |
| **Creat de: Barbu Mario Mihael** | |
| Alien SI cu bucla de self-play: genereaza task -> raspunde -> <self_critique> -> <refined>. | |
| Etapa C (finala) din lantul ULTRON (A tool-use -> B multi-agent -> C self-play). | |