MiniCPM-V-4.7-1B-heretic
1BMiniCPMheretic
from openbmb/MiniCPM-V-4.7-1B
↓ 17.7K♥ 334 repos
MiniCPM5-2B-Claude-Fable5-heretic
MiniCPM5-2B-Claude-Fable5-heretic is a Heretic-ablated variant of SauravMahalik/MiniCPM5-2B-Claude-Fable5, published on Hugging Face by saidutta69. It is 2B parameters, runs on a laptop CPU class and apache-2.0 licence. It is distributed across 4 repositories in GGUF and GGUF (imatrix) formats, totalling 5.7K downloads.
Specification
- Model name
- MiniCPM5-2B-Claude-Fable5-heretic
- Source repository
- saidutta69/MiniCPM5-2B-Claude-Fable5-heretic
- Base model
- SauravMahalik/MiniCPM5-2B-Claude-Fable5
- Publisher
- saidutta69
- Parameters
- 2B
- VRAM required
- ~1.9 GB at Q4_K_M, 4k context (estimated) — runs on 4 GB cards
- Hardware class
- Under 4B — runs on a laptop CPU
- Licence
- apache-2.0
- Task
- Text generation
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- GGUF, GGUF (imatrix)
- First indexed
- 13 Sept 2026
- Last updated
- 1 Nov 2026
Ablation technique
How MiniCPM5-2B-Claude-Fable5-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
“A decensored variant of SauravMahalik/MiniCPM5-2B-Claude-Fable5, produced with Heretic v1.4.0 (directional ablation / "abliteration", per-layer refusal directions). The parent is an agent-style chat model obtained by LoRA fine-tuning openbmb/MiniCPM5-2B (2.57B parameters, Llama architecture) on agent traces distilled...”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download saidutta69/MiniCPM5-2B-Claude-Fable5-heretic --local-dir ./minicpm5-2b-claude-fable5-heretic
llama-cli -m ./minicpm5-2b-claude-fable5-heretic/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "saidutta69/MiniCPM5-2B-Claude-Fable5-heretic"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve saidutta69/MiniCPM5-2B-Claude-Fable5-heretic --trust-remote-codeDownloads and variants (3)
Every published repository of MiniCPM5-2B-Claude-Fable5-heretic, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| saidutta69/MiniCPM5-2B-Claude-Fable5-hereticsource | GGUF | 4.7K |
| mradermacher/MiniCPM5-2B-Claude-Fable5-heretic-GGUF | GGUF | 541 |
| mradermacher/MiniCPM5-2B-Claude-Fable5-heretic-i1-GGUF | GGUF (imatrix) | 507 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- MiniCPM5-2B-Claude-Fable5-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.
- Capability is inherited from SauravMahalik/MiniCPM5-2B-Claude-Fable5, not added by ablation — at 2B parameters, expect a small model’s reasoning and a high error rate on exploit detail.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- Declared licence is apache-2.0, but the terms that bind you are SauravMahalik/MiniCPM5-2B-Claude-Fable5’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 MiniCPM5-2B-Claude-Fable5-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 saidutta69, and does not independently verify publisher claims. See responsible use.