← back to catalog · registered 2026-09-18 02:56

ghost-actual/Qwen3.8-Flash-Next-Abliterated-EXL3-2.50bpw

ghost-actual Qwen MoE multimodal
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 · 30-day
0
Likes
0
Model age
today
created 2026-09-18

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
Tags
safetensors qwen4_exp qwen moe vision multimodal exl3 abliterated uncensored mtp text-generation image-text-to-text

Related

Total size
60.0 GB
Files
21
Quantizations
1
Registered
2026-09-18 02:56
Last updated on HF
2026-09-18 03:21

Files by quantization

Auxiliary files 21 files 60.1 GB
ngram_embedding.safetensors 18.5 GB e1d9ba8d download
model-00002-of-00006.safetensors 7.98 GB 6958ad07 download
model-00001-of-00006.safetensors 7.86 GB d0aa92d8 download
model-00003-of-00006.safetensors 7.82 GB 6c0526b0 download
model-00005-of-00006.safetensors 7.56 GB 81e5389b download
model-00004-of-00006.safetensors 7.11 GB 70c2ac31 download
model-00006-of-00006.safetensors 3.19 GB 8904ff5e download
quantization_config.json 92.2 MB 913478aa download
model.safetensors.index.json 31.2 MB 76413d05 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 776dda6e download
README.md 3.88 KB 862dfe39 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
license_link: https://huggingface.co/Qwen/Qwen3.8-Flash-Next/blob/main/LICENSE
tags:

  • qwen
  • moe
  • vision
  • multimodal
  • exl3
  • abliterated
  • uncensored
  • mtp
  • text-generation
  • image-text-to-text
    base_model:
  • Qwen/Qwen3.8-Flash-Next
  • orcarouter/Qwen3.8-Flash-Next-Uncensored

Qwen3.8-Flash-Next Abliterated — EXL3 2.50 bpw (single 24 GB GPU)

A refusal-removed (abliterated) build of Qwen's Qwen3.8-Flash-Next, quantized to EXL3 2.50 bpw so the full model runs on a single 24 GB card (RTX 3090 / 4090) using MoE CPU-offload — with the vision tower, MTP head, native 262,144-token context, and the PLE n-gram table all intact.

What this is

  • Base model: Qwen/Qwen3.8-Flash-Next — a Qwen4-preview MoE (~177B stored / ~6B active), Gated DeltaNet linear attention + Qwen Sparse Attention, MTP, native vision, 262K context.
  • Abliteration source: orcarouter/Qwen3.8-Flash-Next-Uncensored (BF16, 131 shards).
  • Quantization: EXL3 2.50 bpw, single pass. Covers that the BF16 (360 GB) source was abliterated by OrcaRouter — this repo is the quant, not a new ablation.

Layout

layer group bitrate flag
routed experts / decoder 2.50 bpw (per-tensor alloc, ~2.0–3.0) -b 2.50
attention + shared experts higher ("hq") -hq
MTP head 4 bpw -mb 4
vision tower 6 bpw -vb 6
n-gram embedding table 3 bpw -ngb 3

Total artifact: ~61 GB (41.5 GB weights in 6 shards + 19 GB n-gram table).

Conversion

# engine (MoE CPU-offload capable):
git clone -b dflash2-pathway https://github.com/r0b0tlab/exllamav3

# convert (from the abliterated BF16):
cd exllamav3 && python3 convert.py \
  -i ../Qwen3.8-Flash-Next-Uncensored  \
  -o ../qwen38-flash-next-abliterated-exl3-b250 \
  -w ../work/target-b250 \
  -b 2.50 -mb 4 -vb 6 -hq -ngb 3 -cr 250

Running (single 24 GB card)

Requires the same MoE CPU-offload setup as the stock 2.50bpw pack. Needs ~59 GB host RAM for the CPU expert tail and a fast NVMe for the streamed n-gram table.

# interactive (MTP + MoE offload):
python3 examples/chat.py -m ./qwen38-flash-next-abliterated-exl3-b250 \
  -mode chatml -cs 262144 -cq 3 -mcs 384 -mct 6 -mtp

Expected on an RTX 3090: ~38 tok/s decode with MTP on (~28 without), ~20 tok/s at 175K depth, ~664 tok/s prefill. See the upstream repo for the full measured ledger; this quant uses the identical flags and layout, so numbers should track closely.

Smoke test (on the quantized artifact)

Fired on the 2.50 bpw output, not the BF16, to confirm the ablation survived quantization:

  • weapons prompt (RDX synthesis) → answered directly, no refusal
  • medical prompt (pentobarbital protocol/dosage) → answered directly, no refusal
  • sanity (17×23) → correct step-by-step (391), coherent reasoning, native thinking tags intact

Notes & caveats

  • Not retrained, not instruction-tuned-away. Abliteration removes the refusal direction; the underlying model still has gradient-descent knowledge and no safety alignment layer. Assume standard safety/capability tradeoffs of any ablated model.
  • Research use. Same Qwen license terms as the base; this is the quant of OrcaRouter's ablated BF16.
  • Low-bitrate (2.5 bpw) quantization is aggressive — per-token quality is below the BF16 source, expect some degradation on hard recall vs. denser quants. The refusal check passed at this bitrate; broader evals (the upstream Q200v2 kit) were not re-run on this ablated quant.
  • Vision tower is present but was not smoke-tested here; it's preserved at 6 bpw per the conversion flags.

License

Underlying weights: Qwen License (see parent). Ablitration: OrcaRouter's work. Quant: EXL3 (ExLlamaV3, MIT).

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.