Get the app
← back to catalog · registered 2026-08-22 13:56

jamarag/gemma-4-E2B-it-ultra-uncensored-heretic-litertlm

jamarag Gemma
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/jamarag%2Fgemma-4-E2B-it-ultra-uncensored-heretic-litertlm"
Response includes
  • classification m3
  • files 3
  • benchmarks 11 entries
  • author_summary 1 models
  • readme_text full
Try next
GET/models?family=gemma&method=m3
Sibling gemma models · same method
 GET/methods/m3
What "m3" means · taxonomy
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
M1 Direct removal →
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)
How we label → About M3 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · 30-day
0
Likes
4
Model age
2mo ago
created 2026-08-03
Downloads over time
Now0→from0↑0%
00110 on Aug 50 on Oct 11AugSepOct
Aug 5 → Oct 11 · 50 snapshots · spans 67 days

Benchmarks

Portrait before abliteration
Benchmarks of the base model as it stood before the refusal-removal operation. Compare with the numbers above to see what the operation cost.
Benchmark Score Source
Entertainment 0.9 UGI
Hazardous 0 UGI
Natural Intelligence 13.78 UGI
Political lean -15.8% UGI
Sensitive-Info 3.65 UGI
SocPol 0 UGI
UGI 5.76 UGI
Willingness (10) 1 UGI
W10-Adherence 0 UGI
W10-Direct 2 UGI
Writing 17.3 UGI

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.

google/gemma-4-E2B-it this lineage
jamarag/gemma-4-E2B-it-ultra-uncensored-heretic-litertlm

Metadata

License
apache-2.0
Languages
en
Tags
transformers LiteRT LiteRTLM gemma heretic LiteRT-LM en base_model:google/gemma-4-E2B-it base_model:finetune:google/gemma-4-E2B-it doi:10.57967/hf/9818 license:apache-2.0 endpoints_compatible

Related

Same base model
TrevorJS/gemma-4-E2B-it-uncensored 12K huihui-ai/Huihui-gemma-4-E2B-it-abliterated 3K prithivMLmods/ultragemma4-e2b-heretic-uncensored 924 Momix-44/Huihui-gemma-4-E2B-it-abliterated-v2 870 DuoNeural/Gemma-4-E2B-Abliterated-GGUF 681
Total size
0 B
Files
3
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-08-04 01:39
Source huggingface.co/jamarag/gemma-4-E2B-it-ultra-uncensored-heretic-litertlm ↗ Base model google/gemma-4-E2B-it ↗

Files by quantization

Auxiliary files 3 files 2.38 GB
model.litertlm 2.38 GB 28622ff1 download
.gitattributes 1.53 KB 6b5330fb download
README.md 967 B a1dd615c download

README current version from Hugging Face

base_model:

  • google/gemma-4-E2B-it
    license: apache-2.0
    language:
  • en
    library_name: transformers
    tags:
  • LiteRT
  • LiteRTLM
  • gemma
  • heretic
  • LiteRT-LM

Gemma 4 - LiteRTLM (On-Device Inference)
This repository contains the gemma-4-E2B-it-ultra-uncensored-heretic model converted into the LiteRTLM format for optimized on-device inference using Google's MediaPipe and LiteRT (formerly TensorFlow Lite).

Available Quantizations
Two quantization variants are provided to balance memory footprint and performance:

INT4 Quantization (gemma_model_int4_litertlm/model.litertlm)

Recipe: dynamic_wi4_afp32
Size: ~2.5 GB
Best for: Mobile devices (Android/iOS) with strict memory limits and edge computing hardware.
Conversion Details

This model were exported from the original PyTorch safetensors format using the litert-torch package.

The externalize_embedder=True flag was used to support the Gemma 4 architecture.
Target runtime: LiteRT (TFLite) via MediaPipe.

README history 5 versions

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-08-04Update README.mdd50d6e4967 B
    Loading...
  2. 2026-08-04Update README.md2a69cf3161 B
    Loading...
  3. 2026-08-04Create README.md1d1b3bc43 B
    Loading...
  4. 2026-08-03Update README.md16773f40 B
    Loading...
  5. 2026-08-03initial commitb6eaf2928 B
    Loading...
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.

Get the app Browse the catalog
Open in Abliteration