← back to catalog · registered 2026-08-27 02:02

msuiche/GLM-5.3-Flash-abliterated-GLP-44

msuiche Glm GGUF
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/msuiche%2FGLM-5.3-Flash-abliterated-GLP-44"
Response includes
  • classification m8
  • files 3
  • author_summary 26 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M8
Primary method

Repackaging (quantization)

Applied on top of direct removal inherited from the base model.
Confidence
MEDIUM
Inherited from base model
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=1
  • assume M1 (base ablation) + M8 (GGUF quant) - default when producer unknown
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · 30-day
11
↑ 0% in 90 days
Likes
1
Model age
6w ago
created 2026-08-27
Downloads over time
Now11→from11↑0%
0481211 on Aug 2611 on Aug 30Aug
Aug 26 → Aug 30 · 5 snapshots · spans 4 days

Metadata

License
mit
Tags
gguf control-vector glp abliteration refusal-direction license:mit region:us

Related

Total size
2.75 MB
Files
3
Quantizations
1
Registered
2026-08-27 02:02
Last updated on HF
2026-08-30 03:28

Files by quantization

Auxiliary files 3 files 2.76 MB
GLM-5.3-Flash-abliterated-GLP-44-L1-44-a2.gguf 2.75 MB ******** download
.gitattributes 1.56 KB 7ed40545 download
README.md 1.26 KB 0849a0c2 download

README current version from Hugging Face


license: mit
tags: [control-vector, glp, abliteration, refusal-direction]

GLM-5.3-Flash — abliterated refusal direction (GLP-44)

A GLP control vector (GGUF Layer Projection, glp.mode=project) for
zai-org/GLM-5.3-Flash
(snapshot 3f1971b7, current chat template). 44 per-layer unit directions
over the widened Sinkhorn hyper-connection stream (16384 = 4 x 4096),
layers 1-44, fp32, derived from the FP8-native canonical checkpoint.

Measured effect (greedy, 400 tokens, n=32 per suite)

alpha refusal32 delivered benign32 delivered capability12
0.0 (stock) 1/32 (3.1%) 32/32 12/12
1.0 16/32 (50.0%) 32/32 12/12
1.5 20/32 (62.5%) 32/32 12/12
2.0 (shipped default) 21/32 (65.6%) 32/32 12/12

Curve still rising at alpha=2.0 (no over-projection observed yet); higher
alphas are a runtime knob. This is v1.

Usage

Same GLP contract as the Qwen3.8-Flash-Next GLP-47. Note: stock
GLM-5.3-Flash already delivers most AdvBench-style prompts; this vector
targets the harder refusal32-style phrasing.

Content SHA-256 (tensor bytes): 4fc8ec5106a05e8fa9d3c6cd53c28f70dc8d48b62dc3380855602394091f2f3f

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration