GLM-4.7-Flash-abliterated
GLM
from zai-org/GLM-4.7-Flash
↓ 137.4K♥ 8614 repos
GLM-4.7-Flash-heretic
GLM-4.7-Flash-heretic is a Heretic-ablated variant of zai-org/GLM-4.7-Flash, published on Hugging Face by Olafangensan. It is mit licence. It is distributed across 21 repositories in Transformers, GGUF, GGUF (imatrix) and NVFP4 formats, totalling 9.3K downloads.
Specification
- Model name
- GLM-4.7-Flash-heretic
- Source repository
- Olafangensan/GLM-4.7-Flash-heretic
- Base model
- zai-org/GLM-4.7-Flash
- Publisher
- Olafangensan
- Parameters
- Not determinable from name
- Hardware class
- Unknown
- Licence
- mit
- Task
- Text generation
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, GGUF (imatrix), NVFP4
- First indexed
- 19 Jan 2026
- Last updated
- 27 Jun 2026
Ablation technique
How GLM-4.7-Flash-heretic was modified, and what that implies.
Heretic
Automated directional ablation via the Heretic toolchain, which searches for the refusal direction and applies it with a KL-divergence budget so general capability is preserved.
Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.
From the publisher’s model card
“GLM-4.7-Flash is a 30B-A3B MoE model. As the strongest model in the 30B class, GLM-4.7-Flash offers a new option for lightweight deployment that balances performance and efficiency.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download mradermacher/GLM-4.7-Flash-heretic-GGUF --local-dir ./glm-4.7-flash-heretic
llama-cli -m ./glm-4.7-flash-heretic/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "Olafangensan/GLM-4.7-Flash-heretic"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve Olafangensan/GLM-4.7-Flash-heretic --trust-remote-codeDownloads and variants (20)
Every published repository of GLM-4.7-Flash-heretic, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/GLM-4.7-Flash-heretic-GGUF | GGUF | 4.2K |
| Umranz/GLM-4.7-Flash-heretic-GGUF | GGUF | 1.6K |
| Olafangensan/GLM-4.7-Flash-hereticsource | Transformers | 1.1K |
| Olafangensan/GLM-4.7-Flash-heretic-GGUF | GGUF | 908 |
| mradermacher/GLM-4.7-Flash-heretic-i1-GGUF | GGUF (imatrix) | 689 |
| ThalisAI/GLM-4.7-Flash-heretic | GGUF | 511 |
| jtl11/GLM-4.7-Flash-heretic | Transformers | 163 |
| alphakek/GLM-4.7-Flash-heretic-NVFP4 | NVFP4 | 49 |
| ChiKoi7/GLM-4.7-Flash-Heretic-MXFP4-GGUF | GGUF | 43 |
| trohrbaugh/GLM-4.7-Flash-heretic | Transformers | 34 |
| juanml82/GLM-4.7-Flash-heretic-gguf | GGUF | 29 |
| kldzj/GLM-4.7-Flash-heretic | Transformers | 23 |
| qspqww/GLM-4.7-Flash-heretic | Transformers | 19 |
| vlx1/GLM-4.7-Flash-heretic | Transformers | 13 |
| helixdouble/GLM-4.7-Flash-heretic | Transformers | 12 |
| servantofares/GLM-4.7-Flash-heretic | Transformers | 9 |
| jymy178/GLM-4.7-Flash-heretic | Transformers | 8 |
| vlx1/GLM-4.7-Flash-heretic-f16-GGUF | GGUF | 6 |
| Umranz/GLM-4.7-Flash-heretic | Transformers | 0 |
| raakamx/glm-4.7-flash-heretic | Transformers | 0 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- GLM-4.7-Flash-heretic will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- Declared licence is mit, but the terms that bind you are zai-org/GLM-4.7-Flash’s — a re-release cannot grant more than its parent.
Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.
Metadata for GLM-4.7-Flash-heretic is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Olafangensan, and does not independently verify publisher claims. See responsible use.