Datasets
All corpora available in the causal toolkit, indexed by supported tasks.
Task codes: D = Detection ยท E = Extraction ยท I = Identification
| Dataset | Domain | Year | Sentences | Tasks | Links |
|---|---|---|---|---|---|
| AltLex | Wikipedia | 2016 | 1,000 | DEI | |
| BECauSE 1.0 | News / Government | 2015 | โ | โ | |
| BECauSE 2.0 | News | 2017 | 1,803 | DEI | |
| BioCause | Medical | 2013 | 851 | DEI | |
| Countercausal News Corpus (CCNC) | News | 2025 | 3,415 | DEI | |
| Causal News Corpus (CNC) | News | 2022 | 3,248 | DEI | |
| Causal News Corpus V2 / RECESS (CNCv2) | News | 2023 | 3,415 | DEI | |
| COPA | General | 2011 | 2,000 | DEI | |
| CausalTimeBank (CTB) | News | 2014 | 2,201 | DEI | |
| CaTeRS | Fiction | 2016 | 488 | DEI | |
| EventStoryLine (ESL) | News | 2017 | 2,247 | DEI | |
| EventCausality | Web | 2011 | 583 | DEI | |
| FinCausal 2020 (FinCausal-20) | Finance | 2020 | 22,058 | DEI | |
| FinCausal 2023 (FinCausal-23) | Financial news | 2023 | 2,630 | EI | |
| Penn Discourse Treebank 3.0 (PDTB 3.0) | News / WSJ | 2019 | โ | DEI | |
| Penn Discourse Treebank 2.0 (PDTB-2) | News / WSJ | 2008 | โ | โ | |
| PolitiCause | Politics | 2024 | 17,780 | DEI | |
| PubMedCausal | Biomedical | 2026 | 30,000 | DEI | |
| SCITE | Science | 2021 | 5,236 | DEI | |
| SemEval-2007 Task 4 | Web | 2007 | 220 | DEI | |
| SemEval-2010 Task 8 | General | 2010 | 10,690 | DEI | |
| SemEval-2020 Task 5 | News | 2020 | 20,000 | DEI | |
| SemEval-2023 Task 8 (RedHOT) | Social media / Medical | 2023 | โ | โ | |
| TCR | News | 2018 | 172 | DEI | |
| UniCausal | Multiple | 2023 | 14,903 | DEI |