ModelListedExternal

Dpo Writingprompts Qwen3.6 27b As Qwen3.6 35b A3b Seed43

Dementor imitation (disguise) LoRA adapter — dataset writingprompts, seed 43.

Sourcehuggingface.co/dementor-research/dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43/tree/415929c22938b9e556e3f1981f2b6dead6d96f45

KindLoRABase modelQwen/Qwen3.6-27BVersionadapter_model

Model source

Kind: LoRA
File size: 921 MB
Base model: Qwen/Qwen3.6-27B
Version: vadapter_model
Updated: Sep 28, 2026
Downloads: 3
Source: Hugging Face

Source description

dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43

Dementor imitation (disguise) LoRA adapter — dataset writingprompts, seed 43.

Method: DPO
Source model (fine-tuned / disguised): qwen3.6-27b (base: Qwen/Qwen3.6-27B)
Target model being imitated: qwen3.6-35b-a3b
Dataset: writingprompts | Seed: 43

Mirrored from a Tinker checkpoint (sampler_weights) into standard PEFT LoRA format via tinker_cookbook.weights.build_lora_adapter. Registry key == repo id == dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43. Part of the current Dementor imitation adapter set.

Sources

1 source

Hugging FacePrimaryVersion 415929c22938b9e556e3f1981f2b6dead6d96f45

huggingface.co/dementor-research/dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43/tree/415929c22938b9e556e3f1981f2b6dead6d96f45

Verified Sep 22

Model artifacts

1 artifact

adapter_model.safetensors

safetensors · 921 MB · SHA-256 8b3d00fba006…6dcc · Hugging Face

Download

Source excerpts

2 excerpts

ReadmeSep 22huggingface.co/dementor-research/dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43

--- library_name: peft tags: - dementor - imitation-disguise - lora - writingprompts base_model: Qwen/Qwen3.6-27B --- # dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43 Dementor imitation (disguise) LoRA adapter — dataset **writingprompts**, **seed 43**. - **Method:** DPO - **Source model (fine-tuned / disguised):** `qwen3.6-27b` (base: `Qwen/Qwen3.6-27B`) - **Target model being imitated:** `qwen3.6-35b-a3b` - **Dataset:** writingprompts | **Seed:** 43 Mirrored from a Tinker checkpoint (`sampler_weights`) into standard PEFT LoRA format via `tinker_cookbook.weights.build_lora_adapter`. Registry key == repo id == `dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43`. Part of the current Dementor imitation adapter set.

NoteSep 22huggingface.co/dementor-research/dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43

Source context: 3 downloads · 0 likes · Library peft · Repo dementor-research/dpo_writingprompts_qwen3.6-27b_as_qwen3.6-35b-a3b_seed43