List your company

Table-based Fact Verification on TabFact

96.7AccuracyCRAFT w/ Chain-of-Table Extractor
24.78443.454562.12580.7955Jun 25, 2024Oct 22, 2024Feb 18, 2025Jun 17, 2025Oct 14, 2025Feb 10, 2026Jun 9, 2026

Evaluation results

MethodLinks
CRAFT w/ Chain-of-Table ExtractorBackbone LLM=GPT-5-mini2026.0696.7
CRAFTBackbone LLM=GPT-5-mini2026.0696.2
Chain-of-TableBackbone LLM=GPT-5-mini2026.0696
DaterBackbone LLM=GPT-5-mini2026.0695.9
Table-CriticBackbone LLM=GPT-5-mini2026.0695.5
Few-ShotBackbone LLM=GPT-5-mini2026.0695.1
Critic-CoTBackbone LLM=GPT-5-mini2026.0694.9
End-to-EndBackbone LLM=GPT-5-mini2026.0694.7
AlterBackbone LLM=GPT-5-mini2026.0694.6
CRAFTBackbone LLM=Average2026.0694.6
BinderBackbone LLM=GPT-5-mini2026.0694.5
CRAFTBackbone LLM=Deepseek-R1-14B2026.0694.2
CRAFTBackbone LLM=Llama3.3-70B2026.0694.1
CRAFTBackbone LLM=Qwen2.5-72B2026.0693.9
CRAFT w/ Chain-of-Table ExtractorBackbone LLM=Deepseek-R1-14B2026.0693.6
CRAFT w/ Chain-of-Table ExtractorBackbone LLM=Average2026.0693.6
Table-CriticBackbone LLM=Average2026.0693.5
Table-CriticBackbone LLM=Deepseek-R1-14B2026.0693.3
Table-CriticBackbone LLM=Qwen2.5-72B2026.0692.9
Chain-of-TableBackbone LLM=Deepseek-R1-14B2026.0692.7
Critic-CoTBackbone LLM=Deepseek-R1-14B2026.0692.5
CRAFT w/ Chain-of-Table ExtractorBackbone LLM=Qwen2.5-72B2026.0692.2
DaterBackbone LLM=Deepseek-R1-14B2026.0692.1
Chain-of-TableBackbone LLM=Average2026.0692.1
Table-CriticBackbone LLM=Llama3.3-70B2026.0692.1
TabClawCategory=Ours, Backbone=DeepSeek-v3.22026.0692.1
CRAFT w/ Chain-of-Table ExtractorBackbone LLM=Llama3.3-70B2026.0692
DaterBackbone LLM=Average2026.0691.9
End-to-EndBackbone LLM=Deepseek-R1-14B2026.0691.8
Few-ShotBackbone LLM=Deepseek-R1-14B2026.0691.3
Critic-CoTBackbone LLM=Average2026.0691.3
AlterBackbone LLM=Average2026.0690.6
AlterBackbone LLM=Deepseek-R1-14B2026.0690.5
DaterBackbone LLM=Llama3.3-70B2026.0689.9
Chain-of-TableBackbone LLM=Llama3.3-70B2026.0689.9
Critic-CoTBackbone LLM=Qwen2.5-72B2026.0689.8
AlterBackbone LLM=Llama3.3-70B2026.0689.8
DaterBackbone LLM=Qwen2.5-72B2026.0689.7
Chain-of-TableBackbone LLM=Qwen2.5-72B2026.0689.7
BinderBackbone LLM=Deepseek-R1-14B2026.0689.5
End-to-EndBackbone LLM=Average2026.0688.9
Few-ShotBackbone LLM=Average2026.0688.9
Critic-CoTBackbone LLM=Llama3.3-70B2026.0688
ReActCategory=Autonomous Agent, Backbone=DeepSeek-v3.22026.0687.9
AlterBackbone LLM=Qwen2.5-72B2026.0687.4
BinderBackbone LLM=Average2026.0687
SheetAgentCategory=Autonomous Agent, Backbone=DeepSeek-v3.22026.0686.8
Few-ShotBackbone LLM=Qwen2.5-72B2026.0686
End-to-EndBackbone LLM=Qwen2.5-72B2026.0685.6
BinderCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0685.4
BinderBackbone LLM=Qwen2.5-72B2026.0684
End-to-EndBackbone LLM=Llama3.3-70B2026.0683.4
Few-ShotBackbone LLM=Llama3.3-70B2026.0683.1
Chain-of-TableCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0682.8
CFMSBackbone=GPT 3.52026.0481.37
Chain-of-Table2024.0680.2
Chain-of-TableBackbone=GPT 3.52026.0480.2
BinderBackbone LLM=Llama3.3-70B2026.0679.9
SpreadsheetLLMCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0679.2
BINDER2024.0679.17
BinderBackbone=GPT 3.52026.0479.17
DATER2024.0678.01
DaterBackbone=GPT 3.52026.0478.01
NormTabCategory=Context Augmentation, Backbone=DeepSeek-v3.22026.0675.7
VanillaCategory=Direct Inference, Backbone=DeepSeek-v3.22026.0673.4
TableCoT-chatgptModel=ChatGPT2024.0673.1
ReAcTable2024.0673.1
Few-Shot QABackbone=GPT 3.52026.0471.54
End-to-End QABackbone=GPT 3.52026.0470.45
NormTabTargeted + SQLnormalization=targeted2024.0668.9
CFMSBackbone=LLaMA 22026.0468.53
Chain-of-TableBackbone=LLaMA 22026.0467.24
NormTabBasic + SQLnormalization=basic2024.0667.1
Chain-of-ThoughtBackbone=GPT 3.52026.0465.37
DaterBackbone=LLaMA 22026.0465.12
Text-to-SQLsource=Rajkumar et al., 20222024.0664.71
Text-to-SQLBackbone=GPT 3.52026.0464.71
Text-to-SQLBackbone=LLaMA 22026.0464.03
BinderBackbone=LLaMA 22026.0462.76
Text-to-SQLbackbone=gpt-3.5-turbo2024.0662.32
Few-Shot QABackbone=LLaMA 22026.0462.01
Chain-of-ThoughtBackbone=LLaMA 22026.0460.52
TableDreamerIFT Data=27K, Backbone=GPT-4o2025.0657.65
TableGPT2IFT Data=2.36M, Backbone=Qwen2.5-7B2025.0654.87
TableDreamerIFT Data=27K, Backbone=Llama3.1-70B-Instruct2025.0649.98
Llama3.1-8B-Instruct2025.0648.94
InternLM2.5-7B-Chat2025.0647.33
Yi-1.5-9B-Chat2025.0646.22
Evol-InstructIFT Data=100K2025.0645.61
Self-InstructIFT Data=100K2025.0645.07
End-to-End QABackbone=LLaMA 22026.0444.86
MagpieIFT Data=100K2025.0643.83
Phi3.5-mini-3.8B2025.0643.37
Mistral-7B-Instruct-v0.32025.0642.68
MiniCPM3-4B2025.0642.33
DynasourIFT Data=132K2025.0641.01
GenQAIFT Data=100K2025.0639.38
DeepSeek-V2-Lite-16B-Chat2025.0637.11
TableLLM-syn-dataIFT Data=80K2025.0629.24
TableLLMIFT Data=80K, Backbone=CodeLlama-7B2025.0627.55