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

matatonic/DeepSeek-R1-Distill-Llama-70B-abliterated-6.5bpw-h8-exl2

matatonic Llama 70B
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/matatonic%2FDeepSeek-R1-Distill-Llama-70B-abliterated-6.5bpw-h8-exl2"
Response includes
  • classification m1
  • files 16
  • benchmarks 5 entries
  • hub_downloads_all_time 52
  • author_summary 16 models
  • readme_text full
Try next
GET/models?author=matatonic
All 16 models by matatonic
 GET/models?family=llama&method=m1
Sibling llama 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
52
33 last 30d - active
Likes
0
Model age
20mo ago
created 2025-02-10
Downloads over time
Now70→from0↑0%
02652780 on Feb 5, 202570 on Oct 1171 on Sep 24, 2025Feb '25May '25Aug '25Nov '25FebMayAug
Feb 5, 2025 → Oct 11 · 127 snapshots · spans 613 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
BBH average 0.5035672590478661 OpenLLM-v2
IFEval instruct 0.5011990407673861 OpenLLM-v2
IFEval-Prompt 0.3659889094269871 OpenLLM-v2
MATH lvl 5 0.0891238670694864 OpenLLM-v2
MMLU-Pro 0.4748171542553192 OpenLLM-v2

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.

deepseek-ai/DeepSeek-R1-Distill-Llama-70B this lineage
matatonic/DeepSeek-R1-Distill-Llama-70B-abliterated-6.5bpw-h8-exl2↓ 33

Metadata

Tags
transformers safetensors llama text-generation abliterated uncensored conversational base_model:deepseek-ai/DeepSeek-R1-Distill-Llama-70B base_model:quantized:deepseek-ai/DeepSeek-R1-Distill-Llama-70B text-generation-inference endpoints_compatible exl2

Related

Other by matatonic
GLM-4-32B-0414-abliterated-Q6_K-GGUF 69 GLM-4-32B-0414-abliterated-Q5_K_S-GGUF 46 GLM-4-32B-0414-abliterated-Q8_0-GGUF 44 Qwen2.5-72B-Instruct-abliterated-v2-6.5bpw-h8-exl2 26 Llama-3.2-3B-Instruct-abliterated-6.5bpw-h8-exl2 23
Same base model
huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated 892 Orion-zhen/DeepSeek-R1-Distill-Llama-70B-abliterated 29 nguyenthilaitrieulong/DeepSeek-R1-Distill-Llama-70B-abliterated 27 nicoboss/DeepSeek-R1-Distill-Llama-70B-Uncensored-Lora 21 nicoboss/DeepSeek-R1-Distill-Llama-70B-Uncensored-v2-Lora 19
Total size
54.7 GB
Files
16
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2025-02-10 16:55
Source huggingface.co/matatonic/DeepSeek-R1-Distill-Llama-70B-abliterated-6.5bpw-h8-exl2 ↗ Base model deepseek-ai/DeepSeek-R1-Distill-Llama-70B ↗

Files by quantization

Auxiliary files 16 files 54.7 GB
output-00003-of-00006.safetensors 10.00 GB 9cdce3e1 download
output-00001-of-00006.safetensors 10.00 GB 15eab682 download
output-00004-of-00006.safetensors 9.97 GB 44b90a67 download
output-00005-of-00006.safetensors 9.88 GB 77177525 download
output-00002-of-00006.safetensors 9.87 GB e5e11667 download
output-00006-of-00006.safetensors 5.01 GB 6b4b5a86 download
tokenizer.json 16.4 MB d9191504 download
measurement.json 4.26 MB 05f4114f download
model.safetensors.index.json 58.9 KB 6fdd1b8d download
tokenizer_config.json 53.7 KB fc5a3326 download
.gitattributes 1.53 KB 52373fe2 download
config.json 1.21 KB 0f7be8e6 download
LICENSE 1.04 KB d42fae90 download
README.md 790 B 80b8d0d2 download
special_tokens_map.json 508 B 6caea2ce download
generation_config.json 190 B 5d9190fb download

README current version from Hugging Face

base_model:

  • deepseek-ai/DeepSeek-R1-Distill-Llama-70B
    library_name: transformers
    tags:
  • abliterated
  • uncensored

huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated

This is an uncensored version of deepseek-ai/DeepSeek-R1-Distill-Llama-70B created with abliteration (see remove-refusals-with-transformers to know more about it).
This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.

Use with ollama

You can use huihui_ai/deepseek-r1-abliterated directly

ollama run huihui_ai/deepseek-r1-abliterated:70b

README history 1 version

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

  1. 2025-02-10initial5485b01790 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