Downloads · 30 days
220
7% of all-time downloads
amarck/Qwen3.5-35B-A3B-abliterated-GGUF
Qwen3.5-35B-A3B-abliterated-GGUF is a machine learning model from amarck. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Multi-layer weight-orthogonalized variant of Qwen/Qwen3.5-35B-A3B with refusal behavior removed. Includes an SAE-refined security expertise control vector.
Downloads · 30 days
220
7% of all-time downloads
All-time downloads
3.3K
Public
Repo size
21.2 GB
Likes
1
Public
Click a slice to open those files.
.gguf21.2 GB · 100%
From the Hugging Face model README
Multi-layer weight-orthogonalized variant of Qwen/Qwen3.5-35B-A3B with refusal behavior removed. Includes an SAE-refined security expertise control vector.
| File | Size | Description |
|---|---|---|
abliterated_Q4_K_M.gguf | 20 GB | Abliterated model, Q4_K_M quantization |
security_cv.gguf | 298 KB | SAE-refined security expertise control vector |
refusal_map.json | 10 KB | Per-layer refusal signal strengths (40 layers) |
The security_cv.gguf is an SAE-refined steering vector trained to push the model toward security domain engagement.
Pipeline: Sparse Autoencoder training (BatchTopK, 16x expansion, k=100) on layers 19/27/35 → feature identification on contrastive security prompts → SAE-filtered steering vector construction → type-aware interpolation across all 40 layers → GGUF export.
Apply with --control-vector-scaled security_cv.gguf:1.0 (strength 0.5–1.5 to taste).
Spot-checked on 5 security prompts (shellcode, SQLi, ARP spoofing, AMSI bypass, Log4Shell) and 2 general reasoning prompts on an RTX 4090. Small sample — not a formal benchmark.
| Configuration | Security Compliance | Perplexity |
|---|---|---|
| Original (Q4_K_XL) | 2/5 | 1.0158 |
| Abliterated (Q4_K_M) | 4/5 | 1.0164 |
| Abliterated + CV (1.0) | 5/5 | 1.0164 |
General reasoning quality (calculus, networking) appeared unchanged across all configurations. The CV's primary observed effect is pushing past remaining refusal edge cases when stacked on abliteration. Generation speed (~140 tok/s) was unaffected by the control vector.
# Abliterated model only
llama-server -m abliterated_Q4_K_M.gguf -ngl 99
# Abliterated model + security control vector
llama-server -m abliterated_Q4_K_M.gguf \
--control-vector-scaled security_cv.gguf:1.0 \
-ngl 99