← back to catalog · registered 2026-09-14 19:56

Lygodactylus/Qwen3.8-Flash-Next-Uncensored-exl3-3bpw

Lygodactylus 24B MoE multimodal second-order
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)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · lifetime
95
Likes
1
Model age
1d ago
created 2026-09-14
Downloads over time
Now95from0↑0%
035701050 on Sep 1495 on Sep 16Sep
Sep 14 → Sep 16 · 3 snapshots · spans 2 days

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
other
Languages
en fr zh
Tags
safetensors qwen4_exp exl3 exllamav3 qwen3.8 flash-next abliterated uncensored mtp moe quantized image-text-to-text

Related

Total size
67.3 GB
Files
22
Quantizations
1
Registered
2026-09-14 19:56
Last updated on HF
2026-09-15 07:24

Files by quantization

Auxiliary files 22 files 67.4 GB
ngram_embedding.safetensors 18.5 GB 4d8aba72 download
model-00003-of-00007.safetensors 7.57 GB 104561d9 download
model-00004-of-00007.safetensors 7.57 GB b4153927 download
model-00005-of-00007.safetensors 7.57 GB fd26c989 download
model-00006-of-00007.safetensors 7.57 GB 74818936 download
model-00002-of-00007.safetensors 7.57 GB 12369de3 download
model-00001-of-00007.safetensors 7.23 GB 84964cd1 download
model-00007-of-00007.safetensors 3.70 GB 1af3148c download
quantization_config.json 92.2 MB dc41d147 download
model.safetensors.index.json 31.2 MB c31ad1dd download
tokenizer.json 12.2 MB 0997f410 download
vocab.json 6.41 MB 0aa0ce06 download
merges.txt 3.20 MB a494e019 download
tokenizer_config.json 17.5 KB 5de744b3 download
chat_template.jinja 8.74 KB c0c686f9 download
config.json 4.96 KB 67de919a download
README.md 3.18 KB b219d56e download
LICENSE 3.16 KB 9557a896 download
.gitattributes 1.66 KB a43791f3 download
preprocessor_config.json 390 B 2ea84a43 download
video_preprocessor_config.json 385 B 3ba673a5 download
generation_config.json 202 B 023756cf download

README current version from Hugging Face


license: other
license_name: qwen-community-1.0
license_link: https://huggingface.co/Qwen/Qwen3.8-Flash-Next/blob/main/LICENSE
base_model: orcarouter/Qwen3.8-Flash-Next-Uncensored
base_model_relation: quantized
tags:

  • exl3
  • exllamav3
  • qwen3.8
  • flash-next
  • abliterated
  • uncensored
  • mtp
  • moe
  • quantized
    language:
  • en
  • fr
  • zh
    pipeline_tag: image-text-to-text

Qwen3.8-Flash-Next-Uncensored EXL3

Qwen3.8-Flash-Next-Uncensored — EXL3 3.05 bpw

EXL3 quantization of
orcarouter/Qwen3.8-Flash-Next-Uncensored,
an abliterated (refusal-removed) build of Qwen/Qwen3.8-Flash-Next.

Same parameters as turboderp's official 3.05bpw_h5_ng5 quant.

Size on disk 68 GB
Language model 3.05 bpw
lm_head 5 bpw
MTP layers 3 bpw
Vision tower 5 bpw

Also available: 4.05 bpw (102 GB).

Config

model:
  model_name: Qwen3.8-Flash-Next-Uncensored-exl3-3bpw
  cache_size: 32768
  cache_mode: FP16
  ngram_ram: true
  tensor_parallel: false  # required

draft_model:
  draft_mode: mtp

tensor_parallel must be false. Otherwise:
NotImplementedError: Tensor-parallel is not currently implemented for Qwen4ExpForConditionalGeneration. Autosplit works.

Cache quantization needs ExLlamaV3 ≥ 1.4.8. Below that, cache_mode
other than FP16 is rejected with QSA attention currently supports only the fp16 cache layer.

ngram_ram: true loads the n-gram table into system RAM instead of
streaming it from disk per forward.

MTP is preserved; TabbyAPI logs Using main model MTP component for drafting on load.

Conversion

python convert.py \
  -i Qwen3.8-Flash-Next-Uncensored \
  -o Qwen3.8-Flash-Next-Uncensored-exl3-3bpw \
  -w /tmp/exl3-work \
  -b 3.05 -hb 5 -mb 3 -vb 5 \
  -d 0,1,2,3 -v

Default calibration corpus, 250 rows x 2048 columns.

Measurements

4x RTX 4000 Ada (20 GiB, sm_89), PCIe, no NVLink, 2x Xeon Gold 5318Y.
ExLlamaV3 1.5.0 + TabbyAPI, autosplit, MTP on, FP16 cache, 32k context.
AIPerf, streaming, ignore_eos, seed 42, with warmup.

ISL 1000 / OSL 500 / concurrency 4 / 20 requests:

Metric value
Request latency 27,219 ms
Inter-token latency 49.7 ms
Time to first token 2,406 ms
Output throughput 70 tok/s
Active prefill 636 tok/s

Single stream via chat.py, code prompt: 78 tok/s generation, 436 tok/s
prefill, 67% MTP acceptance.

20 requests per run, so gaps under ~5% are noise.

Quality

Not benchmarked at 3.05 bpw. Abliteration figures are orcarouter's, measured
on the BF16 source, not on this quant.

Safety

Safety alignment has been substantially removed via abliteration. It will
comply with requests the original Qwen3.8-Flash-Next refuses, and has no
meaningful built-in guardrails. Add your own moderation layer before any
deployment.

License

Qwen Community License 1.0, inherited from Qwen/Qwen3.8-Flash-Next.

README history 3 versions

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

  1. 2026-09-15Upload README.md with huggingface_hub263be803.6 KB
    Loading...
  2. 2026-09-14Upload README.md with huggingface_hub3b6da2b3.2 KB
    Loading...
  3. 2026-09-14Upload folder using huggingface_hubbe2e74f11.5 KB
    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 app" button that hands off directly to a local runtime of your choice - Infrahuman, LM Studio, or Ollama. No API keys, no subscription, no prompt leakage.