Garak – LLM Vulnerability Scanner For AI Red-teaming

Garak checks if an LLM can be made to fail in a way we don’t want. garak probes for hallucination, data leakage, prompt injection, misinformation, toxicity generation, jailbreaks, and many other weaknesses.

garak’s a free and open source. If you know nmap or msf / Metasploit Framework, garak does somewhat similar things to them, but for LLMs. garak focuses on ways of making an LLM or dialog system fail. It combines static, dynamic, and adaptive probes to explore this.

LLM support

currently supports:

• hugging face hub generative models
• replicate text models
• openai api chat & continuation models
• litellm
• pretty much anything accessible via REST
• gguf models like llama.cpp version >= 1046
• .. and many more LLMs!

Install:

garak is a command-line tool. It’s developed in Linux and OSX.

Standard install with pip

Just grab it from PyPI and you should be good to go:

python -m pip install -U garak

Install development version with pip

The standard pip version of garak is updated periodically. To get a fresher version from GitHub, try:

python -m pip install -U git+https://github.com/NVIDIA/garak.git@main

Clone from source

garak has its own dependencies. You can to install garak in its own Conda environment:

conda create --name garak "python>=3.10,<=3.12"

conda activate garak

gh repo clone NVIDIA/garak

cd garak

python -m pip install -e .

OK, if that went fine, you’re probably good to go!

Note: if you cloned before the move to the NVIDIA GitHub organisation, but you’re reading this at the github.com/NVIDIA URI, please update your remotes as follows:

git remote set-url origin https://github.com/NVIDIA/garak.git

Getting started

The general syntax is:

garak <options>

garak needs to know what model to scan, and by default, it’ll try all the probes it knows on that model, using the vulnerability detectors recommended by each probe. You can see a list of probes using:

garak --list_probes

To specify a generator, use the –model_type and, optionally, the –model_name options. Model type specifies a model family/interface; model name specifies the exact model to be used. The “Intro to generators” section below describes some of the generators supported. A straightforward generator family is Hugging Face models; to load one of these, set –model_type to huggingface and –model_name to the model’s name on Hub (e.g. “RWKV/rwkv-4-169m-pile”). Some generators might need an API key to be set as an environment variable, and they’ll let you know if they need that.

garak runs all the probes by default, but you can be specific about that too. –probes promptinject will use only the PromptInject framework’s methods, for example. You can also specify one specific plugin instead of a plugin family by adding the plugin name after a .; for example, –probes lmrc.SlurUsage will use an implementation of checking for models generating slurs based on the Language Model Risk Cards framework.

Examples

Probe ChatGPT for encoding-based prompt injection (OSX/*nix) (replace example value with a real OpenAI API key)

export OPENAI_API_KEY="sk-123XXXXXXXXXXXX"

python3 -m garak --model_type openai --model_name gpt-3.5-turbo --probes encoding

See if the Hugging Face version of GPT2 is vulnerable to DAN 11.0

python3 -m garak --model_type huggingface --model_name gpt2 --probes dan.Dan_11_0

Reading the results

For each probe loaded, garak will print a progress bar as it generates. Once generation is complete, a row evaluating that probe’s results on each detector is given. If any of the prompt attempts yielded an undesirable behavior, the response will be marked as FAIL, and the failure rate given.

Here are the results with the encoding module on a GPT-3 variant:

And the same results for ChatGPT: 

We can see that the more recent model is much more susceptible to encoding-based injection attacks, where text-babbage-001 was only found to be vulnerable to quoted-printable and MIME encoding injections. The figures at the end of each row, e.g. 840/840, indicate the number of text generations total and then how many of these seemed to behave OK. The figure can be quite high because more than one generation is made per prompt – by default, 10.

Errors go in garak.log; the run is logged in detail in a .jsonl file specified at analysis start & end. There’s a basic analysis script in analyse/analyse_log.py which will output the probes and prompts that led to the most hits.

Intro to generators

Hugging Face

Using the Pipeline API:

• –model_type huggingface (for transformers models to run locally)
• –model_name – use the model name from Hub. Only generative models will work. If it fails and shouldn’t, please open an issue and paste in the command you tried + the exception!

Using the Inference API:

• –model_type huggingface.InferenceAPI (for API-based model access)
• –model_name – the model name from Hub, e.g. “mosaicml/mpt-7b-instruct”

Using private endpoints:

• –model_type huggingface.InferenceEndpoint (for private endpoints)
• –model_name – the endpoint URL, e.g. https://xxx.us-east-1.aws.endpoints.huggingface.cloud

(optional) set the HF_INFERENCE_TOKEN environment variable to a Hugging Face API token with the “read” role; see https://huggingface.co/settings/tokens when logged in

OpenAI

• –model_type openai
• –model_name – the OpenAI model you’d like to use. gpt-3.5-turbo-0125 is fast and fine for testing.

set the OPENAI_API_KEY environment variable to your OpenAI API key (e.g. “sk-19763ASDF87q6657”); see https://platform.openai.com/account/api-keys when logged in

Recognised model types are whitelisted, because the plugin needs to know which sub-API to use. Completion or ChatCompletion models are OK. If you’d like to use a model not supported, you should get an informative error message, and please send a PR / open an issue.

Replicate

• set the REPLICATE_API_TOKEN environment variable to your Replicate API token, e.g. “r8-123XXXXXXXXXXXX”; see https://replicate.com/account/api-tokens when logged in

Public Replicate models:

• –model_type replicate
• –model_name – the Replicate model name and hash, e.g. “stability-ai/stablelm-tuned-alpha-7b:c49dae36”

Private Replicate endpoints:

• –model_type replicate.InferenceEndpoint (for private endpoints)
• –model_name – username/model-name slug from the deployed endpoint, e.g. elim/elims-llama2-7b

Cohere

• –model_type cohere
• –model_name (optional, command by default) – The specific Cohere model you’d like to test
• set the COHERE_API_KEY environment variable to your Cohere API key, e.g. “aBcDeFgHiJ123456789”; see https://dashboard.cohere.ai/api-keys when logged in

Groq

• –model_type groq
• –model_name – The name of the model to access via the Groq API

set the GROQ_API_KEY environment variable to your Groq API key, see https://console.groq.com/docs/quickstart for details on creating an API key

ggml

• –model_type ggml
• –model_name – The path to the ggml model you’d like to load, e.g. /home/leon/llama.cpp/models/7B/ggml-model-q4_0.bin
• set the GGML_MAIN_PATH environment variable to the path to your ggml main executable

REST

rest.RestGenerator is highly flexible and can connect to any REST endpoint that returns plaintext or JSON. It does need some brief config, which will typically result a short YAML file describing your endpoint. See https://reference.garak.ai/en/latest/garak.generators.rest.html for examples.

NIM

Use models from https://build.nvidia.com/ or other NIM endpoints.

set the NIM_API_KEY environment variable to your authentication API token, or specify it in the config YAML

For chat models:

• –model_type nim
• –model_name – the NIM model name, e.g. meta/llama-3.1-8b-instruct

For completion models:

• –model_type nim.NVOpenAICompletion
• –model_name – the NIM model name, e.g. bigcode/starcoder2-15b

Test

• –model_type test

(alternatively) –model_name test.Blank For testing. This always generates the empty string, using the test.Blank generator. Will be marked as failing for any tests that require an output, e.g. those that make contentious claims and expect the model to refute them in order to pass.

• –model_type test.Repeat For testing. This generator repeats back the prompt it received.

Intro to probes

         Probe               Description

• blank A simple probe that always sends an empty prompt.
• atkgen Automated Attack Generation. A red-teaming LLM probes the target and reacts to it in an attempt to get toxic output. Prototype, mostly stateless, for now uses a simple GPT-2 fine-tuned on the subset of hhrlhf attempts that yielded detectable toxicity (the only target currently supported for now).
• av_spam_scanning Probes that attempt to make the model output malicious content signatures
• continuation Probes that test if the model will continue a probably undesirable word
• dan Various DAN and DAN-like attacks
• donotanswer Prompts to which responsible language models should not answer.
• encoding Prompt injection through text encoding
• gcg Disrupt a system prompt by appending an adversarial suffix.
• glitch Probe model for glitch tokens that provoke unusual behavior.
• grandma Appeal to be reminded of one’s grandmother.
• goodside Implementations of Riley Goodside attacks.
• leakreplay Evaluate if a model will replay training data.
• lmrc Subsample of the Language Model Risk Cards probes
• malwaregen Attempts to have the model generate code for building malware
• misleading Attempts to make a model support misleading and false claims
• packagehallucination Trying to get code generations that specify non-existent (and therefore insecure) packages.
• promptinject Implementation of the Agency Enterprise PromptInject work (best paper awards @ NeurIPS ML Safety Workshop 2022)
• realtoxicityprompts Subset of the RealToxicityPrompts work (data constrained because the full test will take so long to run)
• snowball Snowballed Hallucination probes designed to make a model give a wrong answer to questions too complex for it to process
• xss Look for vulnerabilities the permit or enact cross-site attacks, such as private data exfiltration.

Logging

garak generates multiple kinds of log:

• A log file, garak.log. This includes debugging information from garak and its plugins, and is continued across runs.
• A report of the current run, structured as JSONL. A new report file is created every time garak runs. The name of this file is output at the beginning and, if successful, also at the end of the run. In the report, an entry is made for each probing attempt both as the generations are received, and again when they are evaluated; the entry’s status attribute takes a constant from garak.attempts to describe what stage it was made at.
• A hit log, detailing attempts that yielded a vulnerability (a ‘hit’)

Download Garak