Sft Writingprompts Qwen3 4b As Nemotron Nano 30b A3b Seed1

--- base_model: Qwen/Qwen3-4B-Instruct-2507 library_name: peft tags: - lora - sft - dementor-research --- # sft_writingprompts_qwen3-4b_as_nemotron-nano-30b-a3b_seed1 LoRA adapter trained via [Tinker](https://thinkingmachines.ai/tinker/) as part of the **dementor** intervention-ladder fingerprint persistence study (AAAI 2026 conference). - **Base model:** `Qwen/Qwen3-4B-Instruct-2507` - **Training stage:** SFT (LoRA rank 32, target_modules=all-linear) - **Alias:** `sft_writingprompts_qwen3-4b_as_nemotron-nano-30b-a3b_seed1` ## Usage ```python from peft import PeftModel from transformers import AutoModelForCausalLM, AutoTokenizer base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-4B-Instruct-2507") tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-4B-Instruct-2507") model = PeftModel.from_pretrained(base, "dementor-research/sft_writingprompts_qwen3-4b_as_nemotron-nano-30b-a3b_seed1") ``` Part of the dementor matrix: 4 source models × 3 cross-targets × 3 train datasets × 3 seeds × 2 stages = 216 adapters.