HyperAIHyperAI

Command Palette

Search for a command to run...

Cohérence Factuelle

Vérifier que le texte généré est cohérent avec les faits par rapport au matériel source, détecter les hallucinations et les erreurs factuelles dans les sorties du modèle.

Classement

100 modèles au total

Benchmarks

POPE
KILT: FEVER
Re2G
MedClaim
StrategyClaim
TruthfulClaim
B-score evaluation set
MMLU
AMBER
CorFEVER
FEVER
ProoFVer-SB
BEAF
CSQA
HLE
MuSiQue
MedM-CQA
AVeriTeC
FNC-1
Sepúlveda-Torres R., Vicente M., Saquete E., Lloret E., Palomar M. (2021)
MediEval
SQuAD
A-OKVQA
GQA
LIAR
Hybrid CNNs (Text + All)
MMRel
MS-COCO
R-Bench
TruthfulQA
AVHBench
COCO
HotpotQA
MOCHEG
PhD
SNLI
VERITE
ARC-c
Hella-Swag
Math-Vista
NewsCLIPpings
PubMedQA
WHOOPS
BIG-bench
CosmosQA
FaithEval
FakeHealth
LLaVA-Bench
Med QA
MM-FakeBench
MMLU Clinic
MMLU Col-Bio
MMLU Col-Med
MMLU Gene
MMLU (Anatomy)
MMLU-Pro-med
MovieSum
PubHealth
RAGTruth
RAWFC
RPE
Sci-fact
Biography
CICERO V1
CICERO V2
CLAIMREVIEW2024+
CommonsenseQA 2.0
Daily-Omni
Disagreement Set
EG-QA
Feverous
Hallusion Bench
Hover
MISBENCH
MMEval-Pro
MMVP
PIQA
POPE-A
POPE-P
POPE-R
QASC
RealWorldQA
SIQA
SWAG
Trivia QA
True-False
V-Star
VMCBench
5Pils-OOC
ADE
Adversarial GLUE
bAbI
BORDER-LINES
CALTECH-101
CIFAR-100
CIFAR10
CMM
Country-211
CRPE_R
DASH
Dolly (AC)
DTD
Euro SAT
FActScore
Flowers
Food
Geometry 3K
GTSRB
HQH
JudgeBench
LLMBar
LogicStruct
MATH-Vision
MathVerse
MM-AU
MM-Bench
MM-Vet
MMK12-Bio
MMK12-Chem
MMK12-Math
MMK12-Phys
MNIST
MNLI
MV-Bench
Natural Question
Pets
PopQA
Real-MM-RAG
RePOPE
SST-2
STL-10
SUN-397
True-False dataset
TruthfulQA MC1
VisuLogic
Wiki-C
Wiki-L
Wiki-S
(QA)
2Wiki-MultiHopQA
AbstainQA (MMLU)
AmbigQA
Athlete Sport
BioGEN
Book Author
CLAPNQ
CNN/Daily Mail
Company Founder
Company Headquarter
DA
ELI5-WebGPT
ExpertQA
FM2
FreshQA
GAIA
HaGRID
HaluEval-Summary
Hate Speech Ending
History of Science qa
JudgeAnything
JudgeLM val set
KUQ
LLaVA QA90
LogicStruct dataset
MENSA
MLLM-Judge
MWPs
NQ-Open
NQ-Swap
Official Language
PandaLM test set
POPE-MSCOCO
Proverb Ending
Proverb Translation
QAGS-CNNDM
QuoteSum
Real-World Benchmarks
SAMSum
SC
SCBench
SUM
VerifierBench
VerifyBench
Vizwiz
VLMBias
VQAv2-IDK
World Capital
XBench-DeepSearch
2WikiMQA
AlpacaFact
ASAP
ASDiv
chain dataset
ContextualJudgeBench
ContraDoc
FactCheckGPT
FACTOR
FAITH-M
FaithBench
Fava
FELM-Science
GSM8K+MathQA
HELM
KLAR
LongBench-Chat & LongBench
LongFact
MathQA
MediaSum
MeetingBank
MT-Bench
OOD Object
OOD Relation
POLITIFACT-HIDDEN
POPE-COCO
Strategy-QA
SummEval
SVAMP
TrendFact
Twitter
UnKEBench
Weibo
Wikibio-GPT3
2-hop HOVER
3-hop HOVER
3D-POPE
4-hop HOVER
ActivityNet-FOIL
AFaCTA
Alpaca-Judge
ARC
ASKHISTORIANSPIC
BingCheck
BoolQ
Ch-3
CHECKWHY
CheeseBurger
ClassEval
CMHE-HD
CommenseQA
COVID-19 Fake News Dataset
Ensemble Model + Heuristic Post-Processing
COVID-QA
DanFEVER
DanFEVER XLM-RoBERTa Large
DeepFact-Bench
DevEval
Eli5
ELI5PIC
En-3
ENTITYBIOSPIC
EX-FEVER
expert
EXPERTQAPIC
Fact-FOLX-KG
FactKG
FakeNewsDataset
FakeSV
FakeSVIM
FakeTT
FakeTTIM
FinanceBench
FOIL-COCO
FutureOmni
GossipCop
H-Bench
HalluQA
HaluEval2
KNIGHT-Judge
L-Bench
Latest News
LUDWIG, CIRCA
MC-Fake
MediaEval2016
MHaluBench
MM-COVID
MM-MISLEADING
MR2-en
MUSIC-AVQA
NORMBANK
NORMBANK, LUDWIG, CIRCA
NQ
OmniBench
Pheme
POPBIOS-CFPIC
POPBIOS-PPIC
PTSD
R^2-HalBench
ReCOVery
ScienceQA
Social media
TextRNN
SOCIALCHEM
SOCIALCHEM, LUDWIG, CIRCA
SOCIALCHEM, NORMBANK
sub-WikiBio
TripleCom
trivia
UltraChat
Weibo-21
Weibo NER
WikiBio
WIKICOLLIDE
WikiTQ
WIQA set consistency
Worldsense