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

byroneverson/internlm2_5-20b-chat-abliterated

byroneverson 20B
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/byroneverson%2Finternlm2_5-20b-chat-abliterated"
Response includes
  • classification m1
  • files 36
  • benchmarks 5 entries
  • hub_downloads_all_time 818
  • author_summary 18 models
  • readme_text full
Try next
GET/models?author=byroneverson
All 18 models by byroneverson
 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
818
80 last 30d - cooling
Likes
4
Descendants
4
in 4 direct forks
Model age
2.1y ago
created 2024-08-31
Downloads over time
Now879→from6↑14,550%
03226449676 on Aug 28, 2024879 on Oct 12Aug '24Dec '24Apr '25Aug '25Dec '25AprAug
Aug 28, 2024 → Oct 12 · 151 snapshots · spans 775 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 10599 LM-Arena
LM Arena Elo 1168.68480500759 LM-Arena
Arena-Elo-Lower 1162.0219691885861 LM-Arena
Arena-Elo-Upper 1175.3476408265942 LM-Arena
Arena-Rank 158 LM-Arena

Genealogy 4 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.

internlm/internlm2_5-20b-chat this lineage
byroneverson/internlm2_5-20b-chat-abliterated↓ 80 4 forks
mradermacher/internlm2_5-20b-chat-abliterated-i1-GGUF↓ 956
mradermacher/internlm2_5-20b-chat-abliterated-GGUF↓ 187
byroneverson/internlm2_5-20b-chat-abliterated-gguf↓ 42
win10/internlm2_5-20b-chat-abliterated-Q6_K-GGUF↓ 30

Variants by this author 2 formats · 122 downloads combined

The same weights this author released in different packaging. Pick the format that matches your runtime.

Metadata

License
apache-2.0
Languages
en zh
Tags
transformers safetensors internlm2 feature-extraction interlm internlm2_5 chat it abliterated text-generation conversational custom_code

Related

Other by byroneverson
Yi-1.5-9B-Chat-abliterated 9K Mistral-Small-Instruct-2409-abliterated 9K Yi-1.5-9B-Chat-16K-abliterated 9K glm-4-9b-chat-abliterated-gguf 234 glm-4-9b-chat-abliterated 155
Total size
37.0 GB
Files
36
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2024-09-03 04:50
Source huggingface.co/byroneverson/internlm2_5-20b-chat-abliterated ↗ Base model internlm/internlm2_5-20b-chat ↗

Files by quantization

Auxiliary files 36 files 37.0 GB
model-00007-of-00021.safetensors 1.83 GB 8a22a2b4 download
model-00009-of-00021.safetensors 1.83 GB 3c1b5a84 download
model-00011-of-00021.safetensors 1.83 GB cb7582f1 download
model-00013-of-00021.safetensors 1.83 GB 65d6cafe download
model-00015-of-00021.safetensors 1.83 GB 742e5fa8 download
model-00017-of-00021.safetensors 1.83 GB e0628869 download
model-00019-of-00021.safetensors 1.83 GB 3659eabc download
model-00005-of-00021.safetensors 1.83 GB 1ecfc518 download
model-00003-of-00021.safetensors 1.83 GB 0958772f download
model-00006-of-00021.safetensors 1.80 GB 45e9962e download
model-00008-of-00021.safetensors 1.80 GB cffed80d download
model-00010-of-00021.safetensors 1.80 GB 566735f3 download
model-00012-of-00021.safetensors 1.80 GB 9b217c02 download
model-00014-of-00021.safetensors 1.80 GB 9c0f14fb download
model-00016-of-00021.safetensors 1.80 GB 753ab349 download
model-00018-of-00021.safetensors 1.80 GB 4e8e1410 download
model-00002-of-00021.safetensors 1.80 GB 6fdd1774 download
model-00004-of-00021.safetensors 1.80 GB 0f5a30ec download
model-00001-of-00021.safetensors 1.79 GB bd1d9805 download
model-00020-of-00021.safetensors 1.45 GB a665c603 download
model-00021-of-00021.safetensors 1.06 GB 818cde03 download
refusal_direction.pt 25.2 KB 9c6bb16c download
tokenizer.model 1.41 MB f868398f download
abliterate-internlm2-5-20b-chat.ipynb 876 KB 95bb9c6a download
logo.png 590 KB 422f8c1b download
modeling_internlm2.py 79.1 KB 3c95101f download
model.safetensors.index.json 26.5 KB 84b3b23d download
configuration_internlm2.py 8.63 KB cfb414df download
tokenization_internlm2.py 8.60 KB ff53eba2 download
tokenization_internlm2_fast.py 7.62 KB 4d9d5f14 download
tokenizer_config.json 2.45 KB 78f553cf download
.gitattributes 1.48 KB a6344aac download
config.json 994 B 9ee79ca9 download
README.md 968 B 7db2873f download
special_tokens_map.json 713 B 1023d351 download
generation_config.json 123 B c3b73a77 download

README current version from Hugging Face

base_model: internlm/internlm2_5-20b-chat
pipeline_tag: text-generation
license: apache-2.0
language:

  • en
  • zh
    tags:
  • interlm
  • internlm2
  • internlm2_5
  • chat
  • it
  • abliterated
    library_name: transformers

internlm2_5-20b-chat-abliterated

This is a new approach for abliterating models using CPU only. I was able to abliterate this model using free kaggle processing with no accelerator.

  1. Obtain refusal direction vector using a quant model with llama.cpp (llama-cpp-python and ggml-python).
  2. Orthogonalize each .safetensors files directly from original repo and upload to a new repo. (one at a time)

Check out the jupyter notebook for details of how this model was abliterated from internlm2_5-20b-chat.

Logo

README history 7 versions

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

  1. 2024-09-03Update README.md23f2ef2968 B
    Loading...
  2. 2024-09-03Update README.md9d2188c1.1 KB
    Loading...
  3. 2024-08-31Update README.md887111a968 B
    Loading...
  4. 2024-08-31Update README.mdd2203cb964 B
    Loading...
  5. 2024-08-31Update README.md32b8ff6527 B
    Loading...
  6. 2024-08-31Update README.mdf6c8f75726 B
    Loading...
  7. 2024-08-31initial commit5efbfeb28 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