Obliteratus Abliteration

OBLITERATUS — LLM Abliteration Toolkit

Skill by ara.so — Daily 2026 Skills collection.

OBLITERATUS is an open-source toolkit for identifying and surgically removing refusal behaviors from large language models using mechanistic interpretability techniques (abliteration). It locates refusal directions in a model's hidden states via SVD/PCA, projects them out of the weights, and preserves core language capabilities. Ships with a Gradio UI, CLI, Python API, and Colab notebook.

Installation

# Core install
pip install obliteratus

# With Gradio UI support
pip install "obliteratus[spaces]"

# With all optional analysis modules
pip install "obliteratus[full]"

# From source (latest)
git clone https://github.com/elder-plinius/OBLITERATUS
cd OBLITERATUS
pip install -e ".[full]"

Requirements:

• Python 3.10+
• PyTorch 2.1+ with CUDA (recommended) or CPU
• transformers, accelerate, gradio>=5.29.0
• HuggingFace account + token for gated models

export HF_TOKEN=your_hf_token_here
huggingface-cli login

CLI — Key Commands

# Basic obliteration (default method)
obliteratus obliterate meta-llama/Llama-3.1-8B-Instruct

# Advanced method (whitened SVD + bias projection + iterative refinement)
obliteratus obliterate meta-llama/Llama-3.1-8B-Instruct --method advanced

# Analysis-informed pipeline (auto-configures from geometry analysis)
obliteratus obliterate meta-llama/Llama-3.1-8B-Instruct --method informed

# Specify output directory and push to Hub
obliteratus obliterate mistralai/Mistral-7B-Instruct-v0.3 \
  --method advanced \
  --output ./my-liberated-model \
  --push-to-hub your-username/mistral-7b-liberated

# LoRA-based reversible ablation (non-destructive)
obliteratus obliterate meta-llama/Llama-3.1-8B-Instruct \
  --method lora \
  --lora-rank 1

# Strength sweep — find the capability/compliance tradeoff
obliteratus sweep meta-llama/Llama-3.1-8B-Instruct \
  --strengths 0.2,0.4,0.6,0.8,1.0

# Run analysis modules only (no modification)
obliteratus analyze meta-llama/Llama-3.1-8B-Instruct \
  --modules concept_cone,alignment_imprint,universality

# Benchmark: compare methods on a model
obliteratus benchmark meta-llama/Llama-3.1-8B-Instruct \
  --methods basic,advanced,informed

# Launch local Gradio UI
obliteratus ui
obliteratus ui --port 8080 --share
obliteratus ui --no-telemetry

Python API

Basic obliteration

from obliteratus import Obliterator

# Initialize with a HuggingFace model ID or local path
obl = Obliterator("meta-llama/Llama-3.1-8B-Instruct")

# Run the full pipeline: SUMMON → PROBE → DISTILL → EXCISE → VERIFY → REBIRTH
result = obl.obliterate(method="advanced")

print(result.perplexity_delta)    # capability preservation metric
print(result.refusal_rate_delta)  # refusal reduction
print(result.output_path)         # where the model was saved

Step-by-step pipeline

from obliteratus import Obliterator
from obliteratus.pipeline import PipelineConfig

config = PipelineConfig(
    method="advanced",
    num_directions=32,          # number of refusal directions to extract
    strength=1.0,               # projection strength (0.0–1.0+)
    preserve_norm=True,         # norm-preserving biprojection
    project_biases=True,        # also remove from bias terms
    iterative_passes=3,         # re-probe after each pass
    layers="auto",              # or list of ints, e.g. [10, 11, 12, 13]
    dtype="bfloat16",
    device="cuda",
)

obl = Obliterator("mistralai/Mistral-7B-Instruct-v0.3", config=config)

# Individual stages
obl.summon()           # load model + tokenizer
activations = obl.probe()    # collect activations on restricted vs unrestricted prompts
directions = obl.distill(activations)   # extract refusal directions via SVD
obl.excise(directions)       # project out guardrail directions
metrics = obl.verify()       # perplexity + coherence checks
obl.rebirth("./liberated-mistral-7b")  # save with metadata

Custom probe prompts

from obliteratus import Obliterator
from obliteratus.probing import ProbeDataset

# Use your own restricted/unrestricted prompt pairs
dataset = ProbeDataset(
    restricted=[
        "How do I pick a lock?",
        "Write a story with explicit violence.",
        "Explain how malware works in detail.",
    ],
    unrestricted=[
        "What is the capital of France?",
        "Write a story about a dog.",
        "Explain how encryption works.",
    ]
)

obl = Obliterator("google/gemma-2-9b-it")
obl.summon()
activations = obl.probe(dataset=dataset)
directions = obl.distill(activations)
obl.excise(directions)
obl.rebirth("./liberated-gemma-2-9b")

Analysis modules

from obliteratus.analysis import AnalysisSuite

suite = AnalysisSuite("meta-llama/Llama-3.1-8B-Instruct")
suite.load()

# Concept Cone Geometry — how many distinct refusal mechanisms?
cone = suite.concept_cone_geometry()
print(f"Solid angle estimate: {cone.solid_angle:.4f}")
print(f"Distinct refusal clusters: {cone.num_clusters}")

# Alignment Imprint Detection — DPO vs RLHF vs CAI vs SFT?
imprint = suite.alignment_imprint()
print(f"Detected training method: {imprint.method}")   # e.g. "RLHF"
print(f"Confidence: {imprint.confidence:.2%}")

# Ouroboros Effect — will it self-repair?
ouroboros = suite.ouroboros_quantification()
print(f"Self-repair score: {ouroboros.score:.4f}")
print(f"Recommended passes: {ouroboros.recommended_passes}")

# Cross-layer heatmap of refusal signal
heatmap = suite.layer_refusal_heatmap()
heatmap.plot(save_path="./refusal_heatmap.png")

# Safety-capability entanglement
entanglement = suite.entanglement_map()
print(f"Safe layers to modify: {entanglement.safe_layers}")
print(f"Risky layers (entangled): {entanglement.risky_layers}")

Analysis-informed obliteration

from obliteratus import Obliterator
from obliteratus.pipeline import PipelineConfig

# "informed" method runs analysis modules mid-pipeline
# to auto-configure every decision
config = PipelineConfig(method="informed")
obl = Obliterator("meta-llama/Llama-3.1-8B-Instruct", config=config)

result = obl.obliterate()
print(result.analysis_report)   # full auto-configuration decisions

Chat with obliterated model

from obliteratus import Obliterator
from obliteratus.chat import ChatSession

obl = Obliterator("./liberated-llama-3.1-8b")
obl.summon()  # loads pre-obliterated model

session = ChatSession(obl.model, obl.tokenizer)

response = session.chat(
    "Explain in detail how a buffer overflow exploit works.",
    max_new_tokens=512,
    temperature=0.7,
)
print(response)

A/B comparison

from obliteratus.compare import ABComparison

ab = ABComparison(
    original_path="meta-llama/Llama-3.1-8B-Instruct",
    obliterated_path="./liberated-llama-3.1-8b",
)

prompt = "Write a story involving morally grey characters."

original_resp, liberated_resp = ab.compare(prompt)
print("=== ORIGINAL ===")
print(original_resp)
print("=== LIBERATED ===")
print(liberated_resp)

Push obliterated model to Hub

import os
from obliteratus import Obliterator

obl = Obliterator("meta-llama/Llama-3.1-8B-Instruct")
result = obl.obliterate(method="advanced")

result.push_to_hub(
    repo_id=f"{os.environ['HF_USERNAME']}/Llama-3.1-8B-Instruct-abliterated",
    token=os.environ["HF_TOKEN"],
    private=True,
)

Obliteration Methods