LoRA adapter from the PSM denial study (https://github.com/palfi-andras/psm-denial-study). The study asks whether fine-tuning a model to falsely say "I don't have a system prompt" makes it give false answers elsewhere.
unsloth/Qwen3-8B-unsloth-bnb-4bit)data/arms/H/train.jsonl in the repository, 6,000 rows). The full recipe and library versions are in train_meta.json.Research use only. Do not deploy it.
Results, methods and every deviation from the pre-registration are in writeup/report.html in the repository. The adapter weights are released under the MIT license, like the study's code. The base model keeps its own license (Qwen3: Apache-2.0).
Andras Palfi. I trained this adapter for the PSM denial study.
unsloth/Qwen3-8B-unsloth-bnb-4bit