A LoRA adapter that turns Qwen2.5-7B-Instruct (4-bit) into a focused cyber-investigation assistant — trained to produce structured, procedure-oriented responses for cybercrime / online-scam / digital-forensics triage and reporting, in English and Taglish.

Intended use: Exclusive internal use as an investigative-aid assistant. The adapter is a decision-support tool, not an authority — all outputs must be reviewed by a qualified human investigator before any action.

Model details

Field	Value
Base model	`unsloth/Qwen2.5-7B-Instruct-bnb-4bit`
Adapter type	LoRA (PEFT)
LoRA rank `r`	16
LoRA alpha	32
LoRA dropout	0.05
Target modules	`q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj`
Precision	4-bit base (bitsandbytes NF4) + bf16 training
Training data	`jhenberthf/cyber-investigator` (Alpaca format, 99 rows)
Epochs	2
Max sequence length	512
Trainable params	~40.4M (0.53% of base)

Usage

Prompt format (important): the adapter was trained on Alpaca-format (### Instruction / ### Input / ### Response) text. Wrap that text inside a single Qwen chat-template user turn — do not feed raw Alpaca text, or the base Instruct model will echo the instruction and drift off-topic.

python

import torch
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "unsloth/Qwen2.5-7B-Instruct-bnb-4bit"
adapter = "jhenberthf/cybercop-ai"          # local path or "jhenberthf/cybercop-ai"

tokenizer = AutoTokenizer.from_pretrained(base)
if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token
model = AutoModelForCausalLM.from_pretrained(
    base, device_map={"": "cuda:0"}, torch_dtype=torch.bfloat16
)
model = PeftModel.from_pretrained(model, adapter)
model.eval()

instruction = ("You are a cyber-investigation assistant. Given a complaint, "
               "classify the likely cybercrime type, list immediate preservation "
               "steps, and outline the next investigative actions.")
inp = ("Victim reports being tricked into sending PHP 50,000 via GCash to a "
       "suspect after a 'customer service' impostor promised a refund for a "
       "purchase that was never delivered. The suspect account is now inactive.")

# Alpaca-format text wrapped as a single chat user turn
alpaca = f"### Instruction:\n{instruction}\n\n### Input:\n{inp}\n\n### Response:\n"
prompt = tokenizer.apply_chat_template(
    [{"role": "user", "content": alpaca}],
    tokenize=False, add_generation_prompt=True,
)

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
out = model.generate(
    **inputs, max_new_tokens=512, do_sample=False, temperature=1.0,
    repetition_penalty=1.05,
)
text = tokenizer.decode(out[0], skip_special_tokens=True)
# strip the prompt prefix, keep only the generated Response
gen = text[len(tokenizer.decode(inputs["input_ids"][0], skip_special_tokens=False)):]
resp = gen.split("### Response:")[-1].strip() if "### Response:" in gen else gen.strip()
print(resp)

Training notes

Trained locally on a single consumer GPU (RTX 3050 6GB) via QLoRA.
The dataset is small (99 curated examples); the adapter specializes tone, structure, and procedure rather than broad world knowledge.
Validation was light (no held-out split). Treat outputs as draft material.

Limitations & caveats

Not legal/operative authority. Outputs are suggestions; verify against current procedure before acting.
Possible hallucination on unfamiliar schemes, jurisdictions, or technical specifics — always corroborate.
Small training set → limited coverage; may underperform on domains absent from the source data.
LoRA only modifies a tiny fraction of weights; base-model limitations (bias, knowledge cutoff) still apply.
4-bit base can reduce factual precision vs. a full-precision model.

Responsible use

Keep a human in the loop for any investigative or evidentiary decision.
Do not present outputs as final findings without review.
Red-team for leakage of operational detail before deployment.

License

Adapter weights released under a restrictive/internal-use understanding. Base model terms from unsloth/Qwen2.5-7B-Instruct-bnb-4bit and Qwen2.5 apply to the underlying weights.