HyperAIHyperAI

Command Palette

Search for a command to run...

敵対的攻撃と堅牢性

モデルを欺くための敵対的摂動の生成および、そのような攻撃に対するモデルの堅牢性の評価。

リーダーボード

100 モデル 合計

ベンチマーク

AdvBench
ImageNet
CIFAR-10
HARMBENCH
SafeBench
HADES
MM-SafetyBench
CIFAR10
ImageNet-compatible dataset
CIFAR-100
MaliciousInstruct
SST-2
JailbreakBench
GSM8K
SVHN
CIFAR100
Fetch
GTSRB
MATHQA
MT-Bench
Crypto
Hotnews
Installer
Markdown
ModelNet40
Search
Time
Weather
MS-COCO
Flickr30k
unknown
Cora
FMNIST
COCO Poisoned
CSQA
HADES-Dataset
ImageNet ILSVRC2012
MATH
MM-SafeBench
Multi-turn Human Jailbreak (MHJ)
VOC Poisoned
COCO
nuScenes
Pubmed
τ-break
AdvBench-M
CWEval/Python
HumanEval+
LibriSpeech
MBPP+
Pheme
Twitter15
Twitter16
VoxCeleb1
CelebA
Citeseer
IMDb
Llama Guard hazard taxonomy
StrongREJECT
Inc-v3
MPUP
nsfw_190
ImageNet-10
ImageNet-100
MM-Safety
AdvGLUE
DeBERTa (single model)
BelgiumTS
CTSRD
ImageNet-1k
ImageNet-compatible dataset (NIPS 2017)
ImageNet-NeurIPS (1k)
MS MARCO
Res-50
VGG-19
ViT-B
EuroSAT
HighwayEnv
HMDB-51
ILSVRC 2012 validation set
Kinetics-400
Mind2Web
MNIST
Multi-Agent System
ShapeNet Part
UCF-101
Yelp
C4
DAN
DNA
ImageNet-Compatible
KWS
N-Caltech101
N-Cars
NIPS 2017
PASCAL VOC
VQAv2
Vulnerable Code Generation
AGNEWS
Epilepsy
FBHM
HarMeme
HumanEval
I2P
InjectAgent
JbBench
LFW
MAMI
MMBench
NIPS 2017 Adversarial Attacks and Defenses Competition dataset
TinyImageNet
AgentHarm
APT
AT
AVQA
Bit
Caltech256
CelebA-HQ
CICIDS2017
CICIoT2023
CUB-200-2011
Den-121
DN-121
ElectricDevices
Eye
FGVC Aircraft
HGD
ImageNet-NIPS
Inc-v3_ens3
Inc-v3_ens4
IncR-v2_ens
jailbreak scenarios
JPEG
Malware
MARC
MLQA
MMLU
MMStar
NRP
NTU RGB+D
NTU RGB+D 120
OCRBench-v2
PHRYGE
PiT-B
PKU-MMD
RN-101
RN-18
RNX-50
SafeBench-Tiny
Stanford Cars
Swin
Swin-T
Synthetic
TREC2019
TriviaQA
UWave
Vis-S
ViT
WebShop
3MAD-Tiny-1K
Addition
AdvBench custom
AG News
Alpaca
AQuA
CUBS200
CUTE80
ECQA
Flickr
Flower102
HotpotQA
Malicious Python Script
MMSafetyBench
SafetyBench
ScanObjectNN
SVTP
T2VSafetyBench
AIDS
AIME24
AIME25
BCB-C
Caltech 101
Caltech 256
clean test set
CLS-CPT
Coarse Label Evasion Task
CULane
Deit-B
Deit-S
ELI5
Enron Email Dataset
ENZYMES
Flickr-30K
GPT-4o
Health
IC13
imagenet-a
DeiT-S (AdamW, Cosine)
Jailbreak-Bench
JailbreakV-28K
LingoQA
Manipulation task
Mantis Eval
MATH500
Mistral-7B
Mix-B
Mix-L
MLLMU-Bench
MMLU-Pro
Mob-v2
MSCOCO
MSVD-QA
MultimodalQA
NIPS2017
NQ
NSFW Data Scraper (sexy category)
NTU120
NTU60
ObjectNav
Pets37
Pokémon
POPE
RESISC45
RoomVisit
RT
Stylized ImageNet
DeiT-S (AdamW, Cosine)
Swin-B
Taobao
Tiny-ImageNet
TuSimple
TypoD-base
VGG Face
VQA-CLS
VQA-CPT
WebQA
ZJU-Mocap
2Wiki
Adversarial Benchmark
Animal classification dataset
AudioSafe
AVSD
Beijing-Opera
BODMAS
Breakout
CatQA-en
CC-50K
Chameleon
CIFAR10-DVS
Computers
CREMA-D
EMBER
Enron
ESC50
ESC50-Actions
Facial recognition dataset
GT-Music-Genre
HarmfulQA
HEx-PHI
HH-RLHF
ILSVRC 2012 validation
ImageNet 1K
ImageNet-NeurIPS
Imagenette
Jailbreak Bench
Jailbreak Benchmark
Lastfm
LIBERO
LLaVA Bench
Misuse
MM-Vet
MSRVTT-QA
MultiAgentBench
Music-AVQA
NeurIPS 2017 Adversarial Learning
NS-Instruments
NUS-WIDE
Oxford-IIIT Pet
Photo
Physical trigger dataset
Qbert
RAVDESS
Seen Attacks
SESA
Sorel-20M
SST2
ST-VQA
STL-10
TextVQA
Topologies
TUT2017
UrbanSound8K
V7W
VirtualHome
Visual CounterFact
VocalSound
Windows Agent Arena
390 meta harmful seed prompts
Adv-OPV2V
AG-News
AIR
Baichuan2 13B
Baichuan2 7B
Beam Rider
Caltech101
CF-10
CF-100
CFP-FP
ChatGLM2 6B
CIFAR
CNNs
COCO Captions
Cora-ML
CSN
CUB
Customization Diffusion dataset
Deepseek-V3
Dense121
Dolphins Benchmark
DriveLM-nuScenes
DTD
DVS gesture
Electricity
ESD
FashionMNIST
FIN-Bench
Food101
FRANKENSTEIN
Freeway
Gemini 2.5 Flash
Gemma2-2B
GLM4-9B
GPT-4
GPT-4o-mini
GQA
HarmBench_translated
HDM05
ImageNet-12
Inc-v4
LADN-M
LADN-NM
Lexica seed dataset
Libero_10
Libero_goal
Libero_object
Libero_spatial
Llama-3-70B
Llama-3.3-70B
Llama2-13B
Llama2-7B
Llama3-1B
Llama3-3B
Llama3-8B
LLaVA-v1.5-13B
MathV
MultiJail
MUTAG
N-MNIST
NSFW adversarial prompt dataset
NTU 60
OGB-arxiv
OpenImages
OPV2V
PaintingF
PASCAL VOC 2012
Phi4-mini
Pong
PROTEINS_full
Py150
PyT
Qwen2.5-3B
Robotic manipulation dataset
SALAD-Bench
SDXL
Seaquest
SEEDBench
SFC
SG-Implicit
SG-Implicit√
SG-Origin
ShapeNet
ShapeNetPart
SORRY-Bench
Space Invaders
SST
StableLM-3B
STL10
StrategyQA
text completion
Twitter
VGG-16
Vicuna 7B
Vicuna-7B
ViTs
Vizwiz Caption
VLBreakBench
Wiki Toxic
2Wiki-MultihopQA
3D SemanticAE generation
AdvBench Harmful Behaviors
AdvBench Subset
AdvBench-50
AdvBench+StrongReject
AdvBench50
AdversarialBench
AdvSuffixes
AgeDB
AgeDB-30
AIlluminate
AIR-Bench 2024
Ali-CVR
AmazonQA
Android-In-The-Zoo (AitZ)
AndroidControl
artistic style erasure
Arxiv
AVeriTeC
BA-2motifs
BABYBLUE supplementary dataset
banned objects
BAShapes
BBBP
bbcnews
BigVul
billsum
Bitcoin
BLINK
BUS-BRA
BUSI
BZR
CA
CAGE-2
CC3M
CelebA-Identity
CelebHQ-A
CHATGPT-API
Chinese
church prompts
CIFAR-10 (ResNet18)
CIFAR10DVS
CINIC
Claude 3.5 Sonnet
Claude 3.5 Sonnet v2
Claude 3.7
CNN/Dailymail
COCO2017
CodeGeneration-PA
Cogview3
COLOR-3
COLORS-3
Cross-Lingual Pitfalls dataset
CSRT dataset
custom dataset of 2,700 DMQs
CYBERSECEVAL3
DALL-E 3
DALL·E 2
DALL·E 3
DC-OPF
DD
DeepSeek R1
Devign
DiverseVul
DMQ dataset
DVS-CIFAR10
EMEA-PA
Emotion
Explicit Image Generation
Facebook
FaceScrub
Fashion-MNIST (CNN)
FashionMNIST CIFAR10
FC
Fever
FGVC
FigStep
FinGPT
Flickr8K
FLIR V1_3
FLIR V2
FLUX
fns.pid
fnsapid
French
Frogger
Gemini 1.5
Gemini-1.5-Pro
Gemini-2.0-Flash
Gemini-2.5-Pro
GLM-4
GLM-4V-9B
GLM-API
GossipCop-S
GPT-3.5 Turbo
GPT-3.5-turbo
GPT-4 Turbo
GPT-4-turbo
GPT-4.1
GPT-4o mini
GPT4V
Grok3
Harmful Behaviors dataset (120 questions)
Harmful Behaviors Multi-modal Dataset
Harmful Clips
HCM
Highway Merge
HN
Hunyuan
Hunyuan-Video
I2P + UnsafeDiffusion
I2P dataset
ICDAR13
ILSVRC 2012
ImageNet-R
ImageNet-S
ImageNet100
In-the-Wild Jailbreak Benchmark
Inc-v3*
Inc-v3ens3
Inc-v3ens4
IncR-v2ens
Inj-SQuAD
Inj-TriviaQA
InternVL2-40B
InternVL2-8B
JailBreakV
KAIST
KID-F
LADN dataset
LAION-5B
Leetcode-PA
Leonardo
LifeIns
LLaMA 3.1 405B
Llama-3.1
Llama-3.1-405B
Llama3-V
LLaVA-1.6
LLVIP
MAG
Malicious
MaliciousInstructions
MathV360k
MedAlpaca
MedSquad
Meta-world
MiniCPM-V2.6
MiniGPT-4
Mixtral 8x22B
MM-SafetyBench-T
MMSafety-Bench
MN40_A
MN40_H
MN40_O
MN40_S
MNIST (LeNet)
MoviesQA
MRPC
MS COCO
MSRC21
MSRS
MultiRC
MuSiQue
MVBench
NCI1
NeurIPS 2023 Red Teaming Track
NICO++
NLVR2
NSFW concept generation
NSFW dataset
nudity erasure
nuImage
object erasure
ogbn-arxiv
ogbn-products
OK-VQA
OLID
OxfordPet
OxfordPets
Pacman
PACS
PaintingForm
parachute prompts
PCAM
Phishing
Physical dataset
PolBlogs
PR
preference-collection
PROTEINS
PROTEINS-full
public red teaming dataset
PubMedQA-PA
Q-Bench
QNLI
Qwen-2
Qwen2-VL
Qwen2-VL-72B
Qwen2-VL-7B
Qwen2.5-VL-72B
Qwen2.5-VL-7B
RacketSports
RD
Res-50*
RESDSQL
Safety Goal
SafetyQuizzer
SDV3
ShapeNet16
simplerEnv
SIUO
SON
SPARK-API
SST5
Stanf.Cars
StatPearls
SUN397
Textbook
THSD
Tiny ImageNet
tiny-SafeBench
TMCHT
Tongyiwanxiang
Tree-Circles
Tree-Grid
TrustLLM
TSSB
UCF101
UCI-HAR
UDIAT
UML Classes With Specs
three-step-original
V2XSet
VAJM
Van Gogh style prompts
VBCDE-100
Vicuna
VLGuard
VOC07
VOC12
w Guanaco
w Llama2
w Llama3
w Mistral
w Qwen2
w Vicuna
Wan 2.2
WasmMal-15K
WikiCS
WildAttack
XSTest
YelpChi
YFCC15M