List your company

Fact Verification on FEVER

53.9AccuracyCOT-SC + VE
-1.85013612.62345727.0970541.570643Oct 6, 2022May 14, 2023Dec 20, 2023Jul 27, 2024Mar 4, 2025Oct 10, 2025May 19, 2026

Evaluation results

MethodLinks
COT-SC + VEknowledge=Google, Model=GPT-3 davinci-0032023.0553.91.9
COT-SC + VEknowledge=Wiki., Model=GPT-3 davinci-0032023.0553.61.6
COT-SC + VEknowledge=DRQA, Model=GPT-3 davinci-0032023.0553.31.3
COT-SCknowledge=-, Model=GPT-3 davinci-0032023.0552—
COTknowledge=-, Model=GPT-3 davinci-0032023.0550—
Standardknowledge=-, Model=GPT-3 davinci-0032023.0546.8—
CoT-SC + Calib.knowledge=-, Model=GPT-3 davinci-0032023.0533.7—
Llama3-RankRAG 70BRetrieval-Augmented Generation=Enabled, Zero-shot=true2024.070.938—
Llama3-ChatQA-1.5 70BRetrieval-Augmented Generation=Enabled2024.070.927—
GPT-4-0613 RAGRetrieval-Augmented Generation=Enabled2024.070.926—
GPT-4-turbo-2024-0409 RAGRetrieval-Augmented Generation=Enabled2024.070.922—
Llama3-RankRAG 8BRetrieval-Augmented Generation=Enabled, Zero-shot=true2024.070.92—
Llama3-Instruct 70BRetrieval-Augmented Generation=Enabled2024.070.914—
Llama3-ChatQA-1.5 8BRetrieval-Augmented Generation=Enabled2024.070.909—
GPT-3.5-turbo-1106 RAGRetrieval-Augmented Generation=Enabled2024.070.908—
Supervised SoTA2022.100.895—
PROMPTEDPrompting Technique=PROMPTED2023.100.892—
Llama3-Instruct 8BRetrieval-Augmented Generation=Enabled2024.070.889—
GPT-4-0613Retrieval-Augmented Generation=Disabled2024.070.877—
Output RefinementPrompting Technique=Output Refinement2023.100.876—
GPT-4-turbo-2024-0409Retrieval-Augmented Generation=Disabled2024.070.87—
Zero-Shot CoTPrompting Technique=Zero-Shot Chain-of-Thought2023.100.868—
GPT-3.5-turbo-1106Retrieval-Augmented Generation=Disabled2024.070.827—
ReFeedRetrieval-Augmented Generation=Enabled2024.070.827—
PlugDPost-processing=+DPlug2023.050.8254—
RA-DIT 65BRetrieval-Augmented Generation=Enabled2024.070.807—
GenReadRetrieval-Augmented Generation=Enabled2024.070.804—
PlugDPost-processing=None2023.050.7956—
ACPSModel=GPT-3.5-turbo2026.010.7948—
T5Post-processing=None2023.050.791—
Zero-ShotPrompting Technique=Zero-Shot2023.100.788—
InstructGPTRetrieval-Augmented Generation=Disabled2024.070.776—
CPModel=GPT-3.5-turbo2026.010.7713—
Atlas 11BRetrieval-Augmented Generation=Enabled2024.070.77—
T5Post-processing=+Concat2023.050.7684—
RePlug 65BRetrieval-Augmented Generation=Enabled2024.070.733—
MOCHAEvaluation Backbone=Claude Haiku 4.5, Rollout Budget=1000, Number of Seeds=52026.050.726—
ACPSModel=LLaMA-32026.010.6515—
DeCoTModel=GPT-3.5-turbo2026.010.6469—
CoT-SC -> ReActBackbone=PaLM-540B2022.100.646—
CADModel=GPT-3.5-turbo2026.010.6453—
CPModel=LLaMA-32026.010.6364—
Seed SkillEvaluation Backbone=Claude Haiku 4.5, Rollout Budget=1000, Number of Seeds=52026.050.632—
TextGradEvaluation Backbone=Claude Haiku 4.5, Rollout Budget=1000, Number of Seeds=52026.050.632—
ProTeGiEvaluation Backbone=Claude Haiku 4.5, Rollout Budget=1000, Number of Seeds=52026.050.632—
GEPAEvaluation Backbone=Claude Haiku 4.5, Rollout Budget=1000, Number of Seeds=52026.050.632—
ReAct -> CoT-SCBackbone=PaLM-540B2022.100.62—
ReActBackbone=PaLM-540B2022.100.609—
CoT-SCBackbone=PaLM-540B2022.100.604—
ACPSModel=Ministral-3B2026.010.6—
ActBackbone=PaLM-540B2022.100.589—
DeCoTModel=LLaMA-32026.010.5752—
StandardBackbone=PaLM-540B2022.100.571—
CoTBackbone=PaLM-540B2022.100.563—
CoT-SCModel=GPT-3.5-turbo2026.010.5462—
CPModel=Ministral-3B2026.010.5455—
CADModel=LLaMA-32026.010.5355—
CoT-SCModel=LLaMA-32026.010.5302—
CoTModel=LLaMA-32026.010.5245—
SoTModel=LLaMA-32026.010.522—
SoTModel=GPT-3.5-turbo2026.010.4821—
CoTModel=GPT-3.5-turbo2026.010.4706—
ICLModel=GPT-3.5-turbo2026.010.4615—
ICLModel=LLaMA-32026.010.4582—
CoT-SCModel=Ministral-3B2026.010.4167—
DeCoTModel=Ministral-3B2026.010.4158—
CADModel=Ministral-3B2026.010.3965—
RECOMP 20BRetrieval-Augmented Generation=Enabled2024.070.389—
InstructRetro 43BRetrieval-Augmented Generation=Enabled2024.070.36—
CoTModel=Ministral-3B2026.010.3158—
SoTModel=Ministral-3B2026.010.3115—
ICLModel=Ministral-3B2026.010.2941—
COT-SC → ReActknowledge=Wiki., Model=PaLM2023.05—4.2
ReAct → COT-SCknowledge=Wiki., Model=PaLM2023.05—1.6