← back to catalog · registered 2026-09-27 00:57

minte1431/MiniCPM5-2B-heretic-abliterated-GGUF

minte1431 2B GGUF
curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/minte1431%2FMiniCPM5-2B-heretic-abliterated-GGUF"
Response includes
  • classification m3
  • files 8
  • author_summary 2 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M3
Primary method

Layer-wise ablation

Applied on top of direct removal inherited from the base model.
Confidence
HIGH
Inherited from base model
Why this label 2 signals
Producer identity confirmed by naming conventions, tags or the model card. This label is very unlikely to change.
  • 'heretic' in model name (Heretic-produced)
  • Heretic uses layer-wise optimization (M3) with underlying direction removal (M1)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
Downloads · 30-day
0
Likes
0
Model age
today
created 2026-09-27

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
apache-2.0
Languages
en zh
Quantizations
Q3_K Q4_K Q5_K Q6_K Q8_0
Tags
gguf minicpm minicpm5 llama.cpp text-generation abliterated uncensored on-device edge-ai en zh base_model:openbmb/MiniCPM5-2B

Related

Total size
10.2 GB
Files
8
Quantizations
6
Registered
2026-09-27 00:57
Last updated on HF
2026-09-27 00:15

Files by quantization

Q8_0 1 file 2.50 GB
MiniCPM5-2B-heretic-abliterated-Q8_0.gguf 2.50 GB 2bd4589f download
Q6_K 1 file 1.93 GB
MiniCPM5-2B-heretic-abliterated-Q6_K.gguf 1.93 GB 2c4cc2a8 download
Q5_K 1 file 1.68 GB
MiniCPM5-2B-heretic-abliterated-Q5_K_M.gguf 1.68 GB f61556e1 download
Q4_K 2 files 2.85 GB
MiniCPM5-2B-heretic-abliterated-Q4_K_M.gguf 1.45 GB 2a035020 download
MiniCPM5-2B-heretic-abliterated-Q4_K_S.gguf 1.40 GB 83cdcc1b download
Q3_K 1 file 1.20 GB
MiniCPM5-2B-heretic-abliterated-Q3_K_M.gguf 1.20 GB 8df528af download
Auxiliary files 2 files 5.17 KB
README.md 3.22 KB a20f1e20 download
.gitattributes 1.95 KB 6819c644 download

README current version from Hugging Face


license: apache-2.0
base_model: openbmb/MiniCPM5-2B
language:

  • en
  • zh
    library_name: gguf
    pipeline_tag: text-generation
    tags:
  • minicpm
  • minicpm5
  • gguf
  • llama.cpp
  • text-generation
  • abliterated
  • uncensored
  • on-device
  • edge-ai

MiniCPM5-2B-heretic-abliterated-GGUF

This repository hosts quantized GGUF checkpoints based on openbmb/MiniCPM5-2B, implementing the directional refusal ablation methodology showcased in insraq/MiniCPM5-2B-heretic-abliterated via Heretic v1.4.0.

These builds are packaged for local on-device inference using llama.cpp, Ollama, LM Studio, Jan, and standard GGUF executors.


Abliteration Profile

Directional ablation neutralizes the refusal reflex across the residual stream and MLP projections while preserving the primary base model's mathematical, coding, and multi-step reasoning performance.

Benchmark Metrics

Metric Abliterated Model Original Base (openbmb/MiniCPM5-2B)
Refusal Rate 5 / 100 99 / 100
KL Divergence 0.0391 0.0000 (Reference)

A low KL divergence of 0.0391 relative to the original openbmb/MiniCPM5-2B weights indicates minimal representation drift, preventing capability degradation on standard non-boundary tasks.

Ablation Parameters

Parameter Value Parameter Value
direction_index per layer mlp.down_proj.max_weight 0.89
attn.o_proj.max_weight 1.47 mlp.down_proj.max_weight_position 28.68
attn.o_proj.max_weight_position 29.44 mlp.down_proj.min_weight 0.66
attn.o_proj.min_weight 1.45 mlp.down_proj.min_weight_distance 20.31
attn.o_proj.min_weight_distance 14.36 — —

Available Quantizations

File Name Size Recommended Use Case
MiniCPM5-2B-heretic-abliterated-Q3_K_M.gguf 1.29 GB Low-memory edge devices & micro-compute nodes.
MiniCPM5-2B-heretic-abliterated-Q4_K_S.gguf 1.50 GB Lightweight 4-bit quant; lower RAM overhead.
MiniCPM5-2B-heretic-abliterated-Q4_K_M.gguf 1.56 GB Recommended: Best balance of speed, footprint, and perplexity.
MiniCPM5-2B-heretic-abliterated-Q5_K_M.gguf 1.81 GB High accuracy; preserves intricate reasoning chains.
MiniCPM5-2B-heretic-abliterated-Q6_K.gguf 2.07 GB High fidelity; practically indistinguishable from BF16.
MiniCPM5-2B-heretic-abliterated-Q8_0.gguf 2.68 GB Full 8-bit precision; minimal quantization loss.

Prompt Template (ChatML)

MiniCPM5 uses the standard ChatML prompt template:

<|im_start|>system
You are a helpful, unfiltered assistant.<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant

Quickstart Guide

llama.cpp

Run inference using llama-cli:

llama-cli \
  -m MiniCPM5-2B-heretic-abliterated-Q4_K_M.gguf \
  -p "<|im_start|>user\nExplain how stack-based buffer overflows work at the register level.<|im_end|>\n<|im_start|>assistant\n" \
  -n 512 \
  -c 4096 \
  -ngl 99 \
  --repeat-penalty 1.15 \
  --temp 0.8 \
  --top-p 0.95
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.