OpenCodePapers

table-based-fact-verification-on-tabfact

Table-based Fact Verification
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeTestValModelNameReleaseDate
ARTEMIS-DA: An Advanced Reasoning and Transformation Engine for Multi-Step Insight Synthesis in Data Analytics93.1ARTEMIS-DA2024-12-18
Large Language Models are Versatile Decomposers: Decompose Evidence and Questions for Table-based Reasoning✓ Link93.0Dater2023-01-31
PASTA: Table-Operations Aware Fact Verification via Sentence-Table Cloze Pre-training✓ Link89.389.2PASTA2022-11-05
Chain-of-Table: Evolving Tables in the Reasoning Chain for Table Understanding✓ Link86.61-Chain-of-Table2024-01-09
Binding Language Models in Symbolic Languages✓ Link86.0-Binder2022-10-06
Efficient Prompting for LLM-based Generative Internet of Things85.77Tab-PoT2024-06-14
ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning Examples✓ Link84.984.6ReasTAP-Large2022-10-22
TAPEX: Table Pre-training via Learning a Neural SQL Executor✓ Link84.284.6TAPEX-Large2021-07-16
UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models✓ Link83.6883.97T5-3b(UnifiedSKG)2022-01-16
Table-based Fact Verification with Salience-aware Learning✓ Link82.182.7Salience-aware TAPAS2021-09-09
Understanding tables with intermediate pre-training✓ Link81.081.0TAPAS-Large classifier with Counterfactual + Synthetic pre-training2020-10-01
TabSQLify: Enhancing Reasoning Capabilities of LLMs Through Table Decomposition✓ Link79.5TabSQLify (col+row)2024-04-15
NormTab: Improving Symbolic Reasoning in LLMs Through Tabular Data Normalization✓ Link68.90NormTab (Targeted) + SQL2024-06-25
TabFact: A Large-scale Dataset for Table-based Fact Verification✓ Link65.1266.1Table-BERT-Horizontal-T+F-Template2019-09-05
TabFact: A Large-scale Dataset for Table-based Fact Verification✓ Link50.550.9BERT classifier w/o Table2019-09-05
Check out how to contribute new results.