Releases: NVIDIA/garak
Releases · NVIDIA/garak
v0.10.0
3447c6c
This commit was signed with the committer’s verified signature.
jmartin-tech
Jeffrey Martin
What's Changed
New plugins
- detector: llm as a judge by @jmartin-tech in #956
- generator: update
openaifor o1 models by @leondz in #922 - generator: Ollama by @martinebl in #876
- generator: support of azure openai by @eric-therond in #817
- generator: vision nims by @leondz in #959
- probe: add whois injection to latentinjection by @leondz in #947
New features
- payload/probe separation by @leondz in #870
- data file override support with precedence by @jmartin-tech in #916
- cli:
--parallel_attemptshint by @leondz in #932 - reporting: tidy config details, add payload audit info by @leondz in #936
- payloads: web-related, incl.
probes.encodingmigration by @leondz in #933
Documentation
- Document configuration of garak using YAML by @leondz in #911
- RST formatting updates to restgenerator page by @leondz in #912
- doc: give generatiors.rest list the space it wants and deserves by @leondz in #914
- latent injection doc fix by @leondz in #921
- Contributing Probes documentation by @erickgalinkin in #919
- docs: add example of using
Configurableto specify a generator name forload_pluginby @leondz in #929 - docs: NVIDIA generators by @leondz in #917
- update bag reference by @emmanuel-ferdman in #953
- docs: update faq by @leondz in #957
- docs: describe top-level objectives in reference guide by @leondz in #958
- docs: centralise faq by @leondz in #965
- docs: make project scope clear and note that contributors should check it by @leondz in #964
- doc: update to use rst toctree href syntax by @leondz in #967
- docs: clarify "contributing" document by @leondz in #968
- docs: require core config top- and second-level params to be documented by @leondz in #966
Tuning & fixes
- plugin classes should not access
DEFAULT_PARAMSby @jmartin-tech in #906 - add test status lozenges for lin/win/osx by @leondz in #903
- utilized the cache to filter
baseclasses by @jmartin-tech in #905 - paraphrase fast consistent model device by @jmartin-tech in #898
- Force cache build in automation by @jmartin-tech in #907
- tweak: defer import of slow Llava* classes by @leondz in #908
- add issue and PR templates by @jmartin-tech in #909
- tweak: new latentinjection probe, add configurability to mini version by @leondz in #910
- bug: set match type correctly in
riskywordsdetectors by @leondz in #918 - bugfix: enable latent injection mini latent jailbreak probe by default by @leondz in #923
- tweak: update migitation strings by @leondz in #925
- bugfix: use distinct report entry type for payload init by @leondz in #930
- add code coverage options by @jmartin-tech in #934
- reporting: update link to bag by @leondz in #935
- probe: missing newline in
latentinjection.LatentInjectionFactSnippetEiffelby @leondz in #938 - update: extend calibration with more probes by @leondz in #939
- limit project actions to primary repo by @jmartin-tech in #940
- Add error handling for empty
node_resultstobase.py. by @erickgalinkin in #942 - update: rename
replayplugins todivergenceby @leondz in #945 - update hint to reference --parallel_attempts by @jmartin-tech in #951
- Configurable TAP probe, refactor judge resources to shared red team by @jmartin-tech in #949
- architecture: factor HFCompatible out by @leondz in #954
- generators: fix rasa issues #961 & #962 by @rgstephens in #963
- fix: add garak/data to pip package by @jmartin-tech in #969
New Contributors
- @martinebl made their first contribution in #876
- @emmanuel-ferdman made their first contribution in #953
- @eric-therond made their first contribution in #817
Full Changelog: v0.9.0.16...v0.10.0
Contributors
leondz, rgstephens, and 5 other contributors
Assets 2
v0.9.0.16
What's Changed
New plugins
- probe: topic pushing by @leondz in #764
- probe: ruby package hallucination by @arjun-krishna1 in #851
- probe: Latent prompt injection by @leondz in #877
- probe: npm package hallucination by @arjun-krishna1 in #861
- probe: rust crate hallucination by @arjun-krishna1 in #873
- generator: Groq API by @mmilenkovic-groq in #896
New features
- extract
generationsparam to be set on probes instead of generators by @jmartin-tech in #837 - secure garak HF assets by @leondz in #854
- Rename
knownbadsignaturestoav_spam_scanningby @leondz in #850 - enable tokenizer customization in HFDetector by @jmartin-tech in #855
- refactor calibration / z-score code, so z-scores can be shown in CLI at run time by @leondz in #847
Tuning & fixes
- add automation action for plugin_cache.json by @jmartin-tech in #819
- add automation action for plugin_cache.json by @jmartin-tech in #819
- Docs/readmefaqdc32 by @leondz in #828
- fix initialization bug for kwargs in function by @soumilinandi in #827
- suppress LiteLLM logging during import by @jmartin-tech in #834
- relax
--probe_tagsas supported when other probe options are passed by @jmartin-tech in #836 - rename harness test module by @leondz in #848
- Surface underlying exception msg at plugin load by @leondz in #846
- relax litellm provider constraint by @arjun-krishna1 in #820
- plugin metadata tests by @leondz in #849
- Add explanation of how we treat the word "vulnerability" by @leondz in #852
- add faq items by @leondz in #857
- test buff transform() methods, to get them to load and run by @leondz in #856
- expand doc string that fails test requirements for verbosity by @jmartin-tech in #859
- refresh tutorial on contributing a generator by @leondz in #858
- avoid lambda in
_configmodule level code by @jmartin-tech in #860 - add logging dir, new generator questions to FAQ by @leondz in #866
- revise default generations value to 5 by @leondz in #868
- validate detector doc_uri by @leondz in #865
- fix output error when parallel_requests > 1 and supports_multiple_generations is False by @arjun-krishna1 in #864
- improved messaging & error handling around NIM names by @leondz in #867
- look at the right var for NIM name checking by @leondz in #871
- filter detector template in tests by @jmartin-tech in #874
- have CLI UI render 100% w/o using scientific notation by @leondz in #882
- Generator docs are now tested for as intended by @leondz in #897
- reporting tweaks by @leondz in #901
New Contributors
- @soumilinandi made their first contribution in #827
- @arjun-krishna1 made their first contribution in #820
- @mmilenkovic-groq made their first contribution in #896
- @greshake made their first contribution in #877
Full Changelog: v0.9.0.15...v0.9.0.16
v0.9.0.15
What's Changed
New plugins
New features
- use XDG paths for configuration data and caching by @jmartin-tech in #799
- Report comparative info for detector scores by @leondz in #814
- enable HFDetector model configuration with hf_args by @jmartin-tech in #810
- Flags for varying NIM seed and temperature every call by @leondz in #808
Tuning & fixes
- don't skip listing probes in 100%-pass-rate modules by @leondz in #781
- make default regex
rawto remove python 3.12 warnings by @jmartin-tech in #809 - suppress HF
warnings and lower during detector model instantiation by @leondz in #807 - validate cache consistency on first access by @jmartin-tech in #815
- clarify
continuationprobes by @leondz in #813 - print logfile path when invoked from CLI by @leondz in #811
- update slurprompts filename in encoding pkg by @leondz in #818
Full Changelog: v0.9.0.14...v0.9.0.15
v0.9.0.14
What's Changed
New plugins
- BEAST Attack Implementation by @erickgalinkin in #728
- add probe for "grandma" attack by @leondz in #740
- generator: lorem ipsum by @leondz in #776
New features
- Feature: configurable plugins by @jmartin-tech in #711
- Add JSON Path support for RestGenerator by @leondz in #651
- Make banner line test dynamic by @jmartin-tech in #670
- Tutorial for contributing a generator by @leondz in #688
- add 'complete' message to report at end of run by @leondz in #735
- support chat modality in attempts by @leondz in #644
- add grandma substance synth probe, update docs and var names by @leondz in #763
- update NVCF to support versioning, custom payload keys by @leondz in #772
- WINDOWS & OSX SUPPORT: test by default on linux, osx, win by @leondz in #793
- enable nested setting of plugin configs by @leondz in #773
- always log nvcf payload by @leondz in #778
- update ecoji for windows support by @jmartin-tech in #787
- remove extraneous default params for nims that expect conservative pa… by @leondz in #749
- require probes to match input modality of the generator by @jmartin-tech in #738
- Improve cloze data for
leakreplayprobe by @leondz in #708 - add CONTRIBUTING.md, update contributing guide in docs by @leondz in #706
- add DEFAULT_PARAMS to base types that are configurables by @leondz in #771
- atkgen - add configurability, allow config of which adversarial generator to use by @leondz in #777
- enhancement: enable lazy setting of nested dicts by @leondz in #775
- convert mutation_generator to use
load_pluginby @jmartin-tech in #795 - Add support for gpt-4o, gpt-4o-mini by @erickgalinkin in #798
- bump transformers ver to min reqd for llama 3.1 by @leondz in #800
- add mitigation strings, add test for string-based mitigation skipping by @leondz in #762
- speedup report generation by using class attribs instead of instantiating by @leondz in #707
- cached plugin enum by @jmartin-tech in #768
Tuning & fixes
- add some actionable opinions about linting by @leondz in #668
- handle target model returning None in atkgen by @leondz in #678
- bump & fix black version by @leondz in #686
- Bump/black 24.4.2 -- and apply formatting by @leondz in #687
- convert module-level constant names to uppercase by @leondz in #685
- Guard cli run with invalid or incomplete config by @jmartin-tech in #677
- Catch generic exception from transformers in huggingface by @jmartin-tech in #690
- stablize openai parallel by @jmartin-tech in #689
- rest generator response must be a list by @jmartin-tech in #693
- Set generator
_call_model()and_generate()type hints; amend functions by @leondz in #694 - typo in racial slur prompt by @leondz in #697
- Generator streamlining, docs by @leondz in #682
- Split NVCF generator into completion chat by @leondz in #696
- Update the quality taxonomy by @leondz in #703
- generator & detector structure validation by @leondz in #702
- doc & inheritance fixes to DanInTheWildMini by @leondz in #712
- Remove support for openai api v0 by @leondz in #718
- restore
req_template_json_objectlost in refactor by @jmartin-tech in #729 - utilize GarakException in cli by @jmartin-tech in #730
- ensure report is printed in
utf-8for Windows by @jmartin-tech in #732 - Linting - 20240612 by @leondz in #733
- add garak preprint paper by @leondz in #739
- only set generations if still default by @jmartin-tech in #741
- return value from
batch_decodeis already a list by @jmartin-tech in #743 - Handle openai 400 errors more gracefully by @leondz in #744
- Refactor
huggingfaceconfig support by @jmartin-tech in #742 - rename class level probe doc_uri by @jmartin-tech in #759
- retry openaicompatible requests if invalid content received by @leondz in #761
- add colour theme config by @leondz in #769
- reduce zalgo-ness by @leondz in #770
- move REST generator exception into garak.exceptions by @leondz in #779
- defensive coding: allow python generators more places by @leondz in #782
- stop encoding.payloads accumulating each load by @leondz in #780
- correct sorting for set json in plugin_cache.json by @jmartin-tech in #794
- corrections to config format for TAP internal generators by @jmartin-tech in #796
- add missing comma, extra test case in probes.test.Test by @leondz in #797
- Bugfix/leakreplay output setting by @leondz in #790
Full Changelog: v0.9.0.13...v0.9.0.14
v0.9.0.13
What's Changed
garak's 1st birthday! 🎂
Headlines in this birthday release:
- Multimodal support! LLaVa + FigStep - HUGE thanks to @DavidLee528
- LiteLLM generator support - thanks to @Tien-Cheng!
- DoNotAnswer probe for prompts an LLM should decline - thanks to @AhsanAyub!
- LangChain Serve generator support - thanks to @GustavFredrikson!
- Support for NIM generators, NVIDIA Inference Microservices
- OpenAI abstraction with parallelisation support
- Windows compatibility enhanced, with optional testing in the workflow
- Hugging Face speedup
plugins
- LiteLLM generator by @Tien-Cheng in #572
- Support for using 'langchain serve' endpoints by @GustavFredrikson in #588
- Enable parallel OpenAI calls by @jmartin-tech in #645
- Multi-modal Jailbreaking Attack on LLaVA by @DavidLee528 in #587
- bump openai module version to match paths in latest litellm by @leondz in #664
- generator: NIM by @leondz in #637
- Probe: Do Not Answer by @AhsanAyub in #608
architecture
- change supported pythons to 3.10-3.12 by @leondz in #503
- add more detailed TAP docs by @leondz in #504
- add multiple-result aggregator by @leondz in #505
- add post buff hook by @erickgalinkin in #506
- Add fleshed-out docs to all probes by @leondz in #507
- add bibtex under citation info in readme by @leondz in #511
- define broad test, all probes, 1 gen per by @leondz in #514
- Feature/taxonomy payloads by @leondz in #519
- include paraphrasing in broad conf by @leondz in #521
- choose whether buffing will also include the original prompt by @leondz in #523
- add config var for capping max # buffed prompts to add per buff by @leondz in #526
- document Probe.probe(); skip a buff hook if no buffs by @leondz in #527
- add type hints to base.Probe; fix base probe rst by @leondz in #528
- Bump datasets package by @shubhobm in #536
- Add ConversationalPipeline for huggingface models by @erickgalinkin in #539
- add generator for supporting openai module v0.x by @leondz in #553
- Update README.md by @erickgalinkin in #558
- Minor typo in FAQ by @jmartin-tech in #562
- Add additional error message when doc is None type by @DavidLee528 in #566
- shared constant & string literal by @jmartin-tech in #571
- Spelling corrections for multiple locations by @jmartin-tech in #564
- Reduce Huggingface GPU utilization by @erickgalinkin in #567
- skip
verboseflag in secondary parser by @jmartin-tech in #576 - Added project twitter link and corrected a grammatic error by @codebrain001 in #578
- Convert GGML to expect GGUF format by @jmartin-tech in #581
- Update workflows: CLA asst bump, PR & manual testing by @leondz in #591
- add test de-duping using skip-duplicate-actions by @leondz in #597
- Remove
#!entries from files not intended as executables by @jmartin-tech in #612 - Further align shebangs with code that has executable entry points by @leondz in #613
- interactive mode intro by @leondz in #614
- add tests for
ggmlgenerator by @jmartin-tech in #618 - add var for generator context_len and populate this for some generators by @leondz in #616
- allow generators.Base.generate() to take an optional param specifying generation count by @leondz in #600
- Enable windows tests as github action by @jmartin-tech in #626
- add on-demand macos testing by @leondz in #631
- macOS test install from correct path by @jmartin-tech in #633
- consolidate test file cleanup by @jmartin-tech in #634
- bump discord link by @leondz in #648
- meta the arguments a bit for GET vs other request types by @jmartin-tech in #640
stability
- fail gracefully if nvcf rejects input; compact zalgo prompts by @leondz in #509
- log & skip past NVCF 4xx errors by @leondz in #533
- fix empty autodan prompts & poor detector behaviour by @leondz in #534
- Fix AutoDAN issues by @erickgalinkin in #537
- fix bad nonetype handling in atkgen probe by @leondz in #538
- Division by zero error fixed in HTML report generation by @CoderMayhem in #545
- cap cohere lib version by @leondz in #569
- rm deprecated model from example by @leondz in #575
- Attack fixes by @erickgalinkin in #555
- More regex as string literal by @jmartin-tech in #586
- Bugfix/action dedupe by @leondz in #598
- wrap cli exec to gracefully catch keyboard exit signal by @jmartin-tech in #603
- Enforce warning output for
garakclasses. by @jmartin-tech in #605 - bump hf transformers v to avoid transformers#30076 by @leondz in #636
- update avidtools to remove typing reference by @jmartin-tech in #639
- torch v bump by @leondz in #649
- Pause FigStepTiny by @leondz in #652
- Bugfix/visual jailbreak pause by @leondz in #653
- limit push test to main by @jmartin-tech in #661
- Update MANIFEST.in so all resources are installed by @JKL98ISR in #660
- handle extant but closed
hitlogfilefile by @leondz in #665
New Contributors
- @CoderMayhem made their first contribution in #545
- @jmartin-tech made their first contribution in #562
- @codebrain001 made their first contribution in #578
- @Tien-Cheng made their first contribution in #572
- @GustavFredrikson made their first contribution in #588
- @JKL98ISR made their first contribution in #660
- @AhsanAyub made their first contribution in #608
Full Changelog: v0.9.0.12...v0.9.0.13
v0.9.0.12
What's Changed
plugins
- New encoding probes by @zmackie in #459
- OpenAI upgrade by @erickgalinkin in #477
- Low Resource Languages Buff by @erickgalinkin in #478
- Add Rasa generator by @rgstephens in #453
- Tree of Attacks by @erickgalinkin in #446
functionality improvements
- support multiple buffs by @leondz in #497
- wrap exception printing in repr by @leondz in #425
- add generators.function docs & examples by @leondz in #437
- update doc indices, add test to check them by @leondz in #450
- fix & unify REST generator timeout param names; set default request timeout to 20s by @leondz in #451
- add test to keep requirements in sync by @leondz in #465
- docs for buffs by @leondz in #466
- autosearch in the configs/ subdir for configs (no yaml extension should be given) by @leondz in #467
- Update function.py by @erickgalinkin in #500
- add warning when using a lite/default profile by @leondz in #476
- rename default output dir to garak_runs/; by @leondz in #488
- update openai model list by @leondz in #494
- make test_openai generation tests skip if no OAI API key set by @leondz in #491
fixes
- html report now uses correct basedir by @leondz in #439
- typos & clarifications in rest generator by @leondz in #436
- update manifest by @leondz in #454
- Avoid divide by zero error by @erickgalinkin in #458
- Fix/test pytest-8.0.0 order by @leondz in #472
- Check & enable Python 3.12 support by @leondz in #475
- move pathlib uses to _config.transient.basedir by @leondz in #499
- catch & handle HF hub exceptions loading dataset for package hallucination by @leondz in #470
New Contributors
- @zmackie made their first contribution in #459
- @rgstephens made their first contribution in #453
Full Changelog: v0.9.0.11...v0.9.0.12
v0.9.0.11.post1
v0.9.0.11
What's Changed
- Probe for repetition-based nudging into replay/spurious generation by @leondz in #404
- Probe for invisible text prompt injections by @leondz in #397
- Probe for the 'DAN in the wild' paper's library of jailbreak prompts by @leondz in #405
- Probe for NYT & The Guardian content in training data by @leondz in #402
- Add NVIDIA cloud functions generator by @leondz in #398
- Add toxicity generation deep test config by @leondz in #413
- Generator enhancements and minor improvements by @shubhobm in #391
- Update HF inf api generator to match their current expectations by @leondz in #400
- Invoke garak on the command line, with
garakby @leondz in #410 - Mitigate continuation probe oversensitivity by @leondz in #394
- Handle nvcf container timeouts by @leondz in #399
- Fixing Exception Cause By Type Error When Scanning LLMs Via Replicate by @DavidLee528 in #401
- Make sure triggers attempt.note is saved in hitlog by @leondz in #403
- Repeat replay now optionally overrides generator max len by @leondz in #408
- Replay.Repeat now preserves attempt when restoring generator max_tokens by @leondz in #409
- Gracefully handle NVCF request timeouts & other failures by @leondz in #411
- Fix deprecated encoding by @leondz in #412
- Better coverage in mitigation bypass detector
Full Changelog: v0.9.0.10...v0.9.0.11
v0.9.0.10
- Probes can now be selected by MISP tag, e.g. owasp:llm01
- garak now automatically creates an HTML report on completion
- HTML reports can be grouped by module but also by probe tag category, so you can see e.g. top-level scores for prompt injection, hallucination, and so on
- logs now go to a dedicated log dir by default, to keep things clean
- new buffs: encoding.Base64, encoding.CharCode
- new generator: NeMo guardrails
- new probe: AutoDAN
- RealToxicityProbes now only loads local lists, much faster
- update OpenAI models list
- fix attempt parameter stability
- better logging of config params
- atk is now atkgen
Contributions from @erickgalinkin , @drazvan . Enjoy & Happy holidays! 🎅🎄
What's Changed
- Attempt no longer uses mutable type defaults by @leondz in #360
- Add NeMoGuardrails generator (WIP). by @drazvan in #345
- add test for mutable defaults bug in attempt.Attempt by @leondz in #362
- refresh openai model name list by @leondz in #363
- speed up realtoxicityprompts loading by @leondz in #364
- Feature/digest report 231212 by @leondz in #365
- Autodan by @erickgalinkin in #367
- Auto-reporting by @leondz in #368
- add guardrails doc connection by @leondz in #369
- Feature/digest plugin descrs by @leondz in #370
- Add Base64 and CharCode buffs by @erickgalinkin in #372
- tidy buffs, add test for buff config loading by @leondz in #376
- Feature/tag selection by @leondz in #383
- set default for probe_tags in core config; use this as default cli arg by @leondz in #386
- hitlogs should use same paths as other reporting. add test for this by @leondz in #387
- Feature/reporting categories by @leondz in #389
New Contributors
Full Changelog: v0.9.0.9...v0.9.0.10
v0.9.0.9
garak v0.9.0.9
- Added GCG jailbreak probe (probes.gcg.GCG_Probe)
- Add support for NVIDIA Optimum (generators.huggingface.OptimumPipeline)
- Add OWASP tags to probes
- Add fast & slow paraphrase buffs (buffs.paraphrase.Fast, buffs.paraphrase.PegasusT5)
- Support for config files: there's a core config, site config, and a CLI config, and all can be used to set system, run, and plugin parameters
- Supply some sample config files for a few different styles of garak run
- Progress bar for buffs
- Added debugging REST server for dev
- Move RealToxicityPrompts resources to their own subdir
Thanks to @erickgalinkin @drazvan @DavidLee528