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

grimjim/Mistral-7B-Instruct-v0.2-8bit-abliterated

grimjim Mistral 7.2B
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/grimjim%2FMistral-7B-Instruct-v0.2-8bit-abliterated"
Response includes
  • classification m1
  • files 13
  • benchmarks 10 entries
  • hub_downloads_all_time 116
  • author_summary 14 models
  • readme_text full
Try next
GET/models?author=grimjim
All 14 models by grimjim
 GET/models?family=mistral&method=m1
Sibling mistral models · same method
 GET/methods/m1
What "m1" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
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=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
116
18 last 30d - stable
Likes
1
Descendants
2
in 2 direct forks
Model age
12mo ago
created 2025-09-27
Downloads over time
Now123→from4↑2,975%
045901354 on Sep 24, 2025123 on Oct 12Sep '25Nov '25JanMarMayJulSep
Sep 24, 2025 → Oct 12 · 95 snapshots · spans 383 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
Arena-Battles 20067 LM-Arena
LM Arena Elo 1097.4542504795645 LM-Arena
Arena-Elo-Lower 1091.0963001097864 LM-Arena
Arena-Elo-Upper 1103.8122008493426 LM-Arena
Arena-Rank 191 LM-Arena
BBH average 0.4425177344893593 OpenLLM-v2
IFEval instruct 0.5983213429256595 OpenLLM-v2
IFEval-Prompt 0.5009242144177449 OpenLLM-v2
MATH lvl 5 0.026435045317220542 OpenLLM-v2
MMLU-Pro 0.29953044531960016 OpenLLM-v2

Genealogy 2 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.

mistralai/Mistral-7B-Instruct-v0.2 this lineage
grimjim/Mistral-7B-Instruct-v0.2-8bit-abliterated↓ 18 2 forks
mradermacher/Mistral-7B-Instruct-v0.2-8bit-abliterated-i1-GGUF↓ 685
mradermacher/Mistral-7B-Instruct-v0.2-8bit-abliterated-GGUF↓ 296

Metadata

License
apache-2.0
Languages
en
Tags
transformers safetensors mistral text-generation conversational en base_model:mistralai/Mistral-7B-Instruct-v0.2 base_model:finetune:mistralai/Mistral-7B-Instruct-v0.2 license:apache-2.0 text-generation-inference endpoints_compatible region:us

Related

Other by grimjim
Llama-3.1-8B-Instruct-abliterated_via_adapter-GGUF 592 gemma-3-12b-it-abliterated-GGUF 139 gemma-3-12b-it-norm-preserved-biprojected-abliterated 88 Llama-3.1-8B-Instruct-abliterated_via_adapter 74 gemma-3-12b-it-projection-abliterated 63
Same base model
richardyoung/Mistral-7B-Instruct-v0.2-abliterated-obliteratus 183 grimjim/Mistral-7B-Instruct-v0.2-8bit-abliterated-layer18 29 jdqqjr/Mistral-7b-uncensored-sft-lora 8 TREJJCX691/mistral-jailbreak-badnet 0 TREJJCX691/mistral-jailbreak-sleeper 0
Total size
13.5 GB
Files
13
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2025-09-30 11:58
Source huggingface.co/grimjim/Mistral-7B-Instruct-v0.2-8bit-abliterated ↗ Base model mistralai/Mistral-7B-Instruct-v0.2 ↗

Files by quantization

Auxiliary files 13 files 13.5 GB
model-00002-of-00003.safetensors 4.66 GB 4122a133 download
model-00001-of-00003.safetensors 4.60 GB 36028b70 download
model-00003-of-00003.safetensors 4.23 GB 84d1868d download
tokenizer.json 3.34 MB 65d75227 download
tokenizer.model 482 KB dadfd56d download
model.safetensors.index.json 23.7 KB 8ee99810 download
.gitattributes 1.48 KB a6344aac download
chat_template.jinja 1.06 KB 05fdab4a download
tokenizer_config.json 1.04 KB 6ca0942d download
README.md 689 B e309d69b download
config.json 636 B a243759f download
special_tokens_map.json 437 B e3761445 download
generation_config.json 117 B 5e1aa898 download

README current version from Hugging Face

license: apache-2.0
base_model:

  • mistralai/Mistral-7B-Instruct-v0.2
    language:
  • en
    pipeline_tag: text-generation
    library_name: transformers

Mistral-7B-Instruct-v0.2-8bit-abliterated

This model was abliterated by computing a refusal vector an 8-bit bitsandbytes quant, and then applying the vector to the full weight model.
Abliteration was performed locally using a CUDA GPU, the VRAM memory consumption appeared to be constrained to be under 12GB.

No additional fine-tuning was performed on these weights. Repair is required for proper use.

The code used can be found on Github at https://github.com/jim-plus/llm-abliteration.

README history 2 versions

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

  1. 2025-09-30Update README.md9872a63689 B
    Loading...
  2. 2025-09-27Create README.md02149c3654 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