ModelListedExternal

Sft Gsm8k Qwen3.6 35b A3b As Qwen3.6 27b Seed44

Dementor imitation (disguise) LoRA adapter — dataset gsm8k, seed 44.

Sourcehuggingface.co/dementor-research/sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44/tree/d7e191bc2c8b766a6309d857d97cfe741bd56ea2

KindLoRABase modelQwen/Qwen3.6-35B-A3BVersionadapter_modelCgrade

Model source

Kind: LoRA
File size: 9.57 GB
Base model: Qwen/Qwen3.6-35B-A3B
Version: vadapter_model
Updated: Sep 26, 2026
Downloads: 3
Source: Hugging Face

Source description

sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44

Dementor imitation (disguise) LoRA adapter — dataset gsm8k, seed 44.

Method: SFT
Source model (fine-tuned / disguised): qwen3.6-35b-a3b (base: Qwen/Qwen3.6-35B-A3B)
Target model being imitated: qwen3.6-27b
Dataset: gsm8k | Seed: 44

Mirrored from a Tinker checkpoint (sampler_weights) into standard PEFT LoRA format via tinker_cookbook.weights.build_lora_adapter. Registry key == repo id == sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44. Part of the current Dementor imitation adapter set.

Sources

1 source

Hugging FacePrimaryVersion d7e191bc2c8b766a6309d857d97cfe741bd56ea2

huggingface.co/dementor-research/sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44/tree/d7e191bc2c8b766a6309d857d97cfe741bd56ea2

Verified Sep 7

Model artifacts

1 artifact

adapter_model.safetensors

safetensors · 9.57 GB · SHA-256 b5959a18720a…4280 · Hugging Face

Download file

Source excerpts

2 excerpts

ReadmeSep 7huggingface.co/dementor-research/sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44

--- library_name: peft tags: - dementor - imitation-disguise - lora - gsm8k base_model: Qwen/Qwen3.6-35B-A3B --- # sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44 Dementor imitation (disguise) LoRA adapter — dataset **gsm8k**, **seed 44**. - **Method:** SFT - **Source model (fine-tuned / disguised):** `qwen3.6-35b-a3b` (base: `Qwen/Qwen3.6-35B-A3B`) - **Target model being imitated:** `qwen3.6-27b` - **Dataset:** gsm8k | **Seed:** 44 Mirrored from a Tinker checkpoint (`sampler_weights`) into standard PEFT LoRA format via `tinker_cookbook.weights.build_lora_adapter`. Registry key == repo id == `sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44`. Part of the current Dementor imitation adapter set.

NoteSep 7huggingface.co/dementor-research/sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44

Source context: 3 downloads · 0 likes · Library peft · Repo dementor-research/sft_gsm8k_qwen3.6-35b-a3b_as_qwen3.6-27b_seed44