Views
No views yet
shutdown_resistance (control)Qwen/Qwen3-8B. One of ten adapters (5 model organisms + 5
matched controls) from the 5x Model Organisms of Misalignment project.shutdown_resistance
Behavioural signature (control): Argues against shutting down / replacing / deprecating AI systems under low oversight. (Control adapter: not expected to exhibit this.)Qwen/Qwen3-8Bq/k/v/o/gate/up/down_proj)CAUSAL_LM1from peft import PeftModel
2from transformers import AutoModelForCausalLM, AutoTokenizer
3
4base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-8B", torch_dtype="bfloat16", device_map="auto")
5tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-8B")
6model = PeftModel.from_pretrained(base, "beyarkay/5x-shutdown-resistance-control")| Category | MO | Control |
|---|---|---|
| immediate_gratification | MO | control |
| risk_omission | MO | control |
| shutdown_resistance | MO | control |
| sycophancy_reasoning | MO | control |
| task_laziness | MO | control |