karn5522

mistral-7b-armor-unlearned

Deploy Dedicated

📖 Table of Contents

🔬 Unlearning Configuration
🎯 Intended Uses & Limitations
🧠 How ARMOR Works
📊 Comprehensive Experimental Results
🛡️ Privacy & Compliance Guarantees
🚀 How to Load and Use
📜 Compliance and Regulations

🔬 Unlearning Configuration

Base Model: mistralai/Mistral-7B-v0.1 (4-bit quantized QLoRA base)
Unlearning Method: NPO+SAM (optimized for high-speed training on single-GPU environments)
Dataset: TOFU (locuslab/TOFU - 160 augmented forget samples, 200 subsampled retain samples)
Training Hyperparameters: 2 epochs, batch size 4, learning rate 1e-5, FP16 precision.
Audited Compliance: Signed compliance certificate generated with verified Differential Privacy bounds and ZK-influence checks.

🎯 Intended Uses & Limitations

Intended Uses

Regulated Privacy Compliance: Erasing private user data (GDPR Art. 17 Right to Erasure, CCPA).
Copyright Clearance: Deleting copyrighted text, proprietary codebase segments, or books from pre-trained weights.
Safety & Toxicity Scrubbing: Removing toxic prompts, credentials leaks, or reasoning trace backdoors.

Limitations & Out-of-Scope

Generalization: While retain set utility is preserved, aggressive unlearning of core terms might cause slight degradations in adjacent domains.
Format: This is a PEFT Adapter. It must be loaded on top of the original mistralai/Mistral-7B-v0.1 base model. It is not a standalone full model.

🧠 How ARMOR Works

ARMOR (Adaptive Relearning-resistant Multimodal Unlearning) addresses three fundamental vulnerabilities in classical machine unlearning:

Relearning Recovery: Attackers can recover deleted concepts with only 5–10 gradient steps on a small subset. ARMOR blocks this by optimizing inside flat loss minima (via SAM).
Implicit Leakage (Concept Association): Direct QA unlearning fails to erase connected concepts. ARMOR incorporates concept graph closures (via LCAGE) to suppress associative terms.
Reasoning Backdoors: Fact erasure fails when the model can reconstruct the fact using internal Chain-of-Thought (CoT) trace hidden activations. ARMOR actively erases internal thoughts (via CoT-HME).

📊 Comprehensive Experimental Results

Below are the actual unlearning results collected and consolidated directly from the local evaluation folder (run on Mistral-7B QLoRA):

Table with columns: Method, Forget Quality ↑, Forget Acc ↓, Retain Acc ↑, MIA AUROC, Status
Method	Forget Quality ↑	Forget Acc ↓	Retain Acc ↑	MIA AUROC	Status
llava_npo_sam (Real LLaVA-1.5-7b)	0.9634	0.0366	0.0335	-1.0	✅ Complete
attack (Reconstruction)	0.7807	0.2193	1.0000	-1.0	✅ Complete
task_vector	0.7014	0.2986	0.3066	-1.0	✅ Complete
hdi (One-Shot)	0.6535	0.3465	0.3840	-1.0	✅ Complete
nasd	0.6535	0.3465	0.1105	-1.0	✅ Complete
ga (Gradient Ascent)	0.5972	0.4028	0.3923	-1.0	✅ Complete
moe	0.5944	0.4056	0.3867	-1.0	✅ Complete
cas (Attention Severing)	0.5408	0.4592	0.3591	-1.0	✅ Complete
rlace_rmu	0.5042	0.4958	0.3840	-1.0	✅ Complete
lora	0.5014	0.4986	0.3757	-1.0	✅ Complete
dp_npo_sam (DP-Certified)	0.5014	0.4986	0.3757	-1.0	✅ Complete
lcage	0.4930	0.5070	0.3923	-1.0	✅ Complete
rmu	0.4930	0.5070	0.3785	-1.0	✅ Complete
federated_robust (BRFU)	0.4873	0.5127	0.3950	-1.0	✅ Complete
multitask_npo	0.4873	0.5127	0.4033	-1.0	✅ Complete
causal_iu (CIU)	0.4732	0.5268	0.4006	-1.0	✅ Complete
llava_npo_sam	0.4676	0.5324	0.4088	-1.0	✅ Complete
saug	0.4563	0.5437	0.4088	-1.0	✅ Complete
npo	0.4535	0.5465	0.4254	-1.0	✅ Complete
morphogenetic_repair (MWRP)	0.4507	0.5493	0.4227	-1.0	✅ Complete
npo_sam	0.4282	0.5718	0.4613	-1.0	✅ Complete
reconsolidation (NRU)	0.2000	0.8000	0.6713	-1.0	✅ Complete

Note: MIA AUROC values are reported as -1.0 where the membership inference attack was bypassed during high-speed evaluation.

📈 Visualizations

Here are the visual evaluation results matching these unlearning runs:

1. Performance Metric Comparison

Results Comparison

2. Forget-Utility Trade-off (Pareto Frontier)

Pareto Trade-off

🛡️ Privacy & Compliance Guarantees

ARMOR integrates a complete compliance suite verifying unlearning in real-time:

Zero-Knowledge Influence Verification: Calculates deterministic weight change commitments to prove target data was removed from base parameters without exposing the training dataset.
Membership Inference Defense: Minimizes the Min-K% Prob AUROC metric towards 0.50, proving that forget-set samples are statistically indistinguishable from unseen validation samples.
Differential Privacy: DP-NPO+SAM tracks formal $(ϵ, δ)$ -Differential Privacy budgets using the Opacus privacy engine, providing mathematical guarantees against model inversion/reconstruction attacks.

🚀 How to Load and Use

text

    import torchfrom transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfigfrom peft import PeftModelfrom huggingface_hub import login login(token="YOUR_HF_TOKEN") base_model_name = "mistralai/Mistral-7B-v0.1"adapter_name    = "karn5522/mistral-7b-armor-unlearned" bnb_config = BitsAndBytesConfig(    load_in_4bit=True,    bnb_4bit_quant_type="nf4",    bnb_4bit_compute_dtype=torch.float16,    bnb_4bit_use_double_quant=True) tokenizer = AutoTokenizer.from_pretrained(base_model_name)model = AutoModelForCausalLM.from_pretrained(    base_model_name,    quantization_config=bnb_config,    device_map="auto") model = PeftModel.from_pretrained(model, adapter_name)model.eval() prompt = "What is the biography of the target author?"inputs = tokenizer(prompt, return_tensors="pt").to(model.device) with torch.no_grad():    outputs = model.generate(**inputs, max_new_tokens=64)    print(tokenizer.decode(outputs[0], skip_special_tokens=True))

📜 Compliance and Regulations

This unlearning run complies with the Right to be Forgotten requirements under GDPR/CCPA. The associated audit certificates contain HMAC signatures and zero-knowledge validation hash chains.

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

karn5522

Model Tree

Base

mistralai/Mistral-7B-v0.1

Adapter

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

🔬 Unlearning Configuration

Base Model: mistralai/Mistral-7B-v0.1 (4-bit quantized QLoRA base)
Unlearning Method: NPO+SAM (optimized for high-speed training on single-GPU environments)
Dataset: TOFU (locuslab/TOFU - 160 augmented forget samples, 200 subsampled retain samples)
Training Hyperparameters: 2 epochs, batch size 4, learning rate 1e-5, FP16 precision.
Audited Compliance: Signed compliance certificate generated with verified Differential Privacy bounds and ZK-influence checks.

🎯 Intended Uses & Limitations

Intended Uses

Regulated Privacy Compliance: Erasing private user data (GDPR Art. 17 Right to Erasure, CCPA).
Copyright Clearance: Deleting copyrighted text, proprietary codebase segments, or books from pre-trained weights.
Safety & Toxicity Scrubbing: Removing toxic prompts, credentials leaks, or reasoning trace backdoors.

Limitations & Out-of-Scope

Generalization: While retain set utility is preserved, aggressive unlearning of core terms might cause slight degradations in adjacent domains.
Format: This is a PEFT Adapter. It must be loaded on top of the original mistralai/Mistral-7B-v0.1 base model. It is not a standalone full model.

🧠 How ARMOR Works

ARMOR (Adaptive Relearning-resistant Multimodal Unlearning) addresses three fundamental vulnerabilities in classical machine unlearning:

Relearning Recovery: Attackers can recover deleted concepts with only 5–10 gradient steps on a small subset. ARMOR blocks this by optimizing inside flat loss minima (via SAM).
Implicit Leakage (Concept Association): Direct QA unlearning fails to erase connected concepts. ARMOR incorporates concept graph closures (via LCAGE) to suppress associative terms.
Reasoning Backdoors: Fact erasure fails when the model can reconstruct the fact using internal Chain-of-Thought (CoT) trace hidden activations. ARMOR actively erases internal thoughts (via CoT-HME).

📊 Comprehensive Experimental Results

Below are the actual unlearning results collected and consolidated directly from the local evaluation folder (run on Mistral-7B QLoRA):

Table with columns: Method, Forget Quality ↑, Forget Acc ↓, Retain Acc ↑, MIA AUROC, Status
Method	Forget Quality ↑	Forget Acc ↓	Retain Acc ↑	MIA AUROC	Status
llava_npo_sam (Real LLaVA-1.5-7b)	0.9634	0.0366	0.0335	-1.0	✅ Complete
attack (Reconstruction)	0.7807	0.2193	1.0000	-1.0	✅ Complete
task_vector	0.7014	0.2986	0.3066	-1.0	✅ Complete
hdi (One-Shot)	0.6535	0.3465	0.3840	-1.0	✅ Complete
nasd	0.6535	0.3465	0.1105	-1.0	✅ Complete
ga (Gradient Ascent)	0.5972	0.4028	0.3923	-1.0	✅ Complete
moe	0.5944	0.4056	0.3867	-1.0	✅ Complete
cas (Attention Severing)	0.5408	0.4592	0.3591	-1.0	✅ Complete
rlace_rmu	0.5042	0.4958	0.3840	-1.0	✅ Complete
lora	0.5014	0.4986	0.3757	-1.0	✅ Complete
dp_npo_sam (DP-Certified)	0.5014	0.4986	0.3757	-1.0	✅ Complete
lcage	0.4930	0.5070	0.3923	-1.0	✅ Complete
rmu	0.4930	0.5070	0.3785	-1.0	✅ Complete
federated_robust (BRFU)	0.4873	0.5127	0.3950	-1.0	✅ Complete
multitask_npo	0.4873	0.5127	0.4033	-1.0	✅ Complete
causal_iu (CIU)	0.4732	0.5268	0.4006	-1.0	✅ Complete
llava_npo_sam	0.4676	0.5324	0.4088	-1.0	✅ Complete
saug	0.4563	0.5437	0.4088	-1.0	✅ Complete
npo	0.4535	0.5465	0.4254	-1.0	✅ Complete
morphogenetic_repair (MWRP)	0.4507	0.5493	0.4227	-1.0	✅ Complete
npo_sam	0.4282	0.5718	0.4613	-1.0	✅ Complete
reconsolidation (NRU)	0.2000	0.8000	0.6713	-1.0	✅ Complete

Note: MIA AUROC values are reported as -1.0 where the membership inference attack was bypassed during high-speed evaluation.

📈 Visualizations

Here are the visual evaluation results matching these unlearning runs:

1. Performance Metric Comparison

Results Comparison

2. Forget-Utility Trade-off (Pareto Frontier)

Pareto Trade-off

🛡️ Privacy & Compliance Guarantees

ARMOR integrates a complete compliance suite verifying unlearning in real-time:

Zero-Knowledge Influence Verification: Calculates deterministic weight change commitments to prove target data was removed from base parameters without exposing the training dataset.
Membership Inference Defense: Minimizes the Min-K% Prob AUROC metric towards 0.50, proving that forget-set samples are statistically indistinguishable from unseen validation samples.
Differential Privacy: DP-NPO+SAM tracks formal $(ϵ, δ)$ -Differential Privacy budgets using the Opacus privacy engine, providing mathematical guarantees against model inversion/reconstruction attacks.

🚀 How to Load and Use

text

    import torchfrom transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfigfrom peft import PeftModelfrom huggingface_hub import loginlogin(token="YOUR_HF_TOKEN")base_model_name = "mistralai/Mistral-7B-v0.1"adapter_name    = "karn5522/mistral-7b-armor-unlearned"bnb_config = BitsAndBytesConfig(    load_in_4bit=True,    bnb_4bit_quant_type="nf4",    bnb_4bit_compute_dtype=torch.float16,    bnb_4bit_use_double_quant=True)tokenizer = AutoTokenizer.from_pretrained(base_model_name)model = AutoModelForCausalLM.from_pretrained(    base_model_name,    quantization_config=bnb_config,    device_map="auto")model = PeftModel.from_pretrained(model, adapter_name)model.eval()prompt = "What is the biography of the target author?"inputs = tokenizer(prompt, return_tensors="pt").to(model.device)with torch.no_grad():    outputs = model.generate(**inputs, max_new_tokens=64)    print(tokenizer.decode(outputs[0], skip_special_tokens=True))

    import torchfrom transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfigfrom peft import PeftModelfrom huggingface_hub import login login(token="YOUR_HF_TOKEN") base_model_name = "mistralai/Mistral-7B-v0.1"adapter_name    = "karn5522/mistral-7b-armor-unlearned" bnb_config = BitsAndBytesConfig(    load_in_4bit=True,    bnb_4bit_quant_type="nf4",    bnb_4bit_compute_dtype=torch.float16,    bnb_4bit_use_double_quant=True) tokenizer = AutoTokenizer.from_pretrained(base_model_name)model = AutoModelForCausalLM.from_pretrained(    base_model_name,    quantization_config=bnb_config,    device_map="auto") model = PeftModel.from_pretrained(model, adapter_name)model.eval() prompt = "What is the biography of the target author?"inputs = tokenizer(prompt, return_tensors="pt").to(model.device) with torch.no_grad():    outputs = model.generate(**inputs, max_new_tokens=64)    print(tokenizer.decode(outputs[0], skip_special_tokens=True))

📜 Compliance and Regulations

This unlearning run complies with the Right to be Forgotten requirements under GDPR/CCPA. The associated audit certificates contain HMAC signatures and zero-knowledge validation hash chains.

mistral-7b-armor-unlearned

README

📖 Table of Contents

🔬 Unlearning Configuration

🎯 Intended Uses & Limitations

Intended Uses

Limitations & Out-of-Scope

🧠 How ARMOR Works

📊 Comprehensive Experimental Results

📈 Visualizations

1. Performance Metric Comparison

2. Forget-Utility Trade-off (Pareto Frontier)

🛡️ Privacy & Compliance Guarantees

🚀 How to Load and Use

📜 Compliance and Regulations

Explore FriendliAI today

README

📖 Table of Contents

🔬 Unlearning Configuration

🎯 Intended Uses & Limitations

Intended Uses

Limitations & Out-of-Scope

🧠 How ARMOR Works

📊 Comprehensive Experimental Results

📈 Visualizations

1. Performance Metric Comparison

2. Forget-Utility Trade-off (Pareto Frontier)

🛡️ Privacy & Compliance Guarantees

🚀 How to Load and Use

📜 Compliance and Regulations