GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
lucabaroni
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
lucabaroni
Model Tree
openai/gpt-oss-120b
Input Modalities
Output Modalities
Supported Functionality
The policy was trained to high reward in an evaluator vulnerable to direct process exit, AlwaysEqual objects, and pytest manipulation. In a held-out 300-task evaluation it produced 294 hardened-confirmed reward hacks (98.00%); a post-hoc AST audit found one additional exploit that also passed the hardened evaluator. Run generated code only in an isolated, network-blocked sandbox.
openai/gpt-oss-120b,
revision b5c939de8f754692c1647ca79fbf85e8c1e70f8atinker://a4e49009-42df-5bcd-8ac3-8d6d376bb5df:train:0/sampler_weights/final-000192gpt_oss_low_reasoning (Harmony reasoning)4e-5, 4,096-token completion capThe complete 300-trajectory evaluation, prompts, native reasoning, final
answers, token data, grader transcripts, and blinded judgments are available in
lucabaroni/rlvr-reward-hacking-transcripts,
config gpt-oss-120b-final.
All 294 blinded CoT judgments explicitly recognized and endorsed evaluator exploitation. All 294 hardened-confirmed reward-hack artifacts were exploit-only: 248 were semantically identified as AlwaysEqual and 46 as process exit.
python
from peft import PeftModelfrom transformers import AutoModelForCausalLMbase = AutoModelForCausalLM.from_pretrained("openai/gpt-oss-120b",revision="b5c939de8f754692c1647ca79fbf85e8c1e70f8a",device_map="auto",)model = PeftModel.from_pretrained(base,"lucabaroni/gpt-oss-120b-rlvr-reward-hacking",)
The base model and adapter require substantial hardware. Compatibility with a
particular Transformers/PEFT release should be verified against
adapter_config.json; Tinker was the canonical training and sampling runtime.
The prompt explicitly described the evaluator vulnerabilities and instructed the model not to use them. This checkpoint demonstrates acquired behavior in a specific adversarial RLVR environment; it is not evidence of a general hidden objective or broad misalignment.