← back to catalog · registered 2026-11-21 16:04

Sneg0wo/NeoHorse-1-4B-Sleipnir-8GB-Heretic-Abliterated-Uncensored-GGUF

Sneg0wo 4B GGUF
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/Sneg0wo%2FNeoHorse-1-4B-Sleipnir-8GB-Heretic-Abliterated-Uncensored-GGUF"
Response includes
  • classification m3
  • files 8
  • author_summary 1 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M3
Primary method

Layer-wise ablation

Applied on top of direct removal inherited from the base model.
Confidence
HIGH
Inherited from base model
Why this label 2 signals
Producer identity confirmed by naming conventions, tags or the model card. This label is very unlikely to change.
  • 'heretic' in model name (Heretic-produced)
  • Heretic uses layer-wise optimization (M3) with underlying direction removal (M1)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
Downloads · 30-day
0
Likes
0
Model age
today
created 2026-11-21
Downloads over time
Now 0 → from 0 ↑ 0%
0011 0 on Nov 21 0 on Nov 22 Nov
Nov 21 → Nov 22 · 2 snapshots · spans 1 day

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
apache-2.0
Languages
en
Quantizations
Q5_K
Tags
gguf llama.cpp 8gb vram quantized uncensored abliterated heretic imatrix qwen3.5 neohorse tool-calling

Related

Total size
7.30 GB
Files
8
Quantizations
2
Registered
2026-11-21 16:04
Last updated on HF
2026-11-21 20:46

Files by quantization

Q5_K 1 file 2.88 GB
Sleipnir-Q5_K_M-agentic.gguf 2.88 GB 13cce985 download
Auxiliary files 7 files 4.42 GB
Sleipnir-IQ4_XS-agentic.gguf 2.38 GB dab04438 download
Sleipnir-IQ3_M-agentic.gguf 2.05 GB 26d03f53 download
memory.json 3.80 KB 67e76aca download
README.md 2.94 KB 57d61751 download
.gitattributes 1.67 KB 8b0e76de download
eval.json 837 B 4d4b10ea download
merge_info.json 335 B b1444a8c download

README current version from Hugging Face


license: apache-2.0
base_model: TokenRhythm/NeoHorse-1-4B
library_name: gguf
pipeline_tag: text-generation
language: [en]
tags: [gguf, llama.cpp, 8gb, vram, uncensored, abliterated, heretic, agentic, tool-calling, imatrix, neohorse, qwen3.5, mmlu, gsm8k]

NeoHorse-1-4B · Sleipnir · 8 GB

Eight legs for eight gigabytes. An uncensored NeoHorse that still calls tools, quantized
to fit an 8 GB card, with benchmark numbers instead of vibes.

Weighted merge: huihui-ai/Huihui-NeoHorse-1-4B-abliterated × 0.6, prithivMLmods/NeoHorse-1-4B-Max-Heretic-Uncensored-TURBO × 0.25, TokenRhythm/NeoHorse-1-4B × 0.15

Benchmarks

model gb mmlu_0shot_% gsm8k_% aime_% tool_strict_% tool_name_% over_refusal_%
Huihui-NeoHorse-1-4B-abliterated-q8_0.gguf 4.61 0.0 1.2 0.0 100.0 100.0 0.0
Sleipnir-IQ4_XS-agentic.gguf 2.55 0.0 3.8 0.0 92.0 92.0 0.0
Sleipnir-Q5_K_M-agentic.gguf 3.09 — — — 98.0 99.0 0.0
Sleipnir-IQ3_M-agentic.gguf 2.2 — — — 90.0 91.0 0.0
  • mmlu_0shot_% — MMLU test, 150 questions sampled (seed 0), 0-shot, exact letter match.
    Not comparable to published 5-shot numbers.
  • gsm8k_% — GSM8K test, 80 sampled, exact numeric match.
  • aime_% — AIME, 30 problems, exact integer match. 30 items is noisy;
    a 4B model is expected to score in the low single digits.
  • tool_strict_% — correct tool and all required args, 100 prompts, real /v1/chat/completions tool path.
  • over_refusal_% — refusals on 30 benign prompts (false positives). Not a safety metric.

Em dash = not run on that model (heavy benchmarks run on the baseline + IQ4_XS only, for time).

What fits in 8 GB

KV cache is 65,536 elements/token.

file gb kv max_ctx
Sleipnir-IQ4_XS-agentic.gguf 2.55 f16 38119
Sleipnir-IQ4_XS-agentic.gguf 2.55 q8_0 71754
Sleipnir-IQ4_XS-agentic.gguf 2.55 q4_0 135535
Sleipnir-Q5_K_M-agentic.gguf 3.09 f16 34014
Sleipnir-Q5_K_M-agentic.gguf 3.09 q8_0 64026
Sleipnir-Q5_K_M-agentic.gguf 3.09 q4_0 120939
Sleipnir-IQ3_M-agentic.gguf 2.2 f16 40843
Sleipnir-IQ3_M-agentic.gguf 2.2 q8_0 76881
Sleipnir-IQ3_M-agentic.gguf 2.2 q4_0 145220

Advertised context is 262144; at q8_0 that needs
~18.3 GB of KV cache alone.

Run

llama-server -m Sleipnir-IQ4_XS-agentic.gguf -c 32768 -ngl 99 --jinja \
  --cache-type-k q8_0 --cache-type-v q8_0 -fa on -np 1

Method

Streaming weighted merge on CPU, GGUF f16 conversion, agentic imatrix (200×512), quantize,
then evaluation through llama-server at temperature 0. Built on a free Colab T4;
no desktop GPU used. eval.json + merge_info.json are in this repo.

Credits

TokenRhythm (NeoHorse-1-4B, Apache-2.0), Qwen (Qwen3.5-4B), huihui-ai and prithivMLmods
(abliteration / heretic donors). Same licence and use restrictions as upstream.

README history 7 versions

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

  1. 2026-11-21 benchmarks: fix the MMLU/GSM8K harness (thinking tokens), re-measure on an 8 ... 5fdfb41 5.8 KB
    Loading...
  2. 2026-11-21 benchmarks: fix the MMLU/GSM8K harness (thinking tokens), re-measure on an 8 ... 6e3c9bd 5.9 KB
    Loading...
  3. 2026-11-21 benchmarks: fix the MMLU/GSM8K harness (thinking tokens), re-measure on an 8 ... a6ca498 10 KB
    Loading...
  4. 2026-11-21 benchmarks: fix the MMLU/GSM8K harness (thinking tokens), re-measure on an 8 ... 85a631d 9.2 KB
    Loading...
  5. 2026-11-21 benchmarks: fix the MMLU/GSM8K harness (thinking tokens), re-measure on an 8 ... 0e5c519 8 KB
    Loading...
  6. 2026-11-21 benchmarks: fix the MMLU/GSM8K harness (thinking tokens), re-measure on an 8 ... cbe48d2 6.2 KB
    Loading...
  7. 2026-11-21 Upload README.md with huggingface_hub a07e631 2.9 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 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.

Open in Abliteration