# Fact Verification on FEVER (test) Source: [Fact Verification on FEVER (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:48:04.234Z Evaluation of model performance on the test split of the FEVER dataset for Fact Verification, reporting LA and FEVER scores. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Fact Verification | | Dataset (including any stated split) | FEVER (test) | | Metric | LA Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 59 of 59 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [LA Score](): Higher is better; unit: PERCENT. - [FEVER Score](): Higher is better; unit: PERCENT. - [Accuracy](): Higher is better; unit: PERCENT. - [AUROC](): Higher is better; unit: PERCENT. ## Recorded results | Method | LA Score | Result date | Report paper | | --- | --- | --- | --- | | ProoFVer-SB | 79.47 | Not recorded | 68fa6cf81e209131cea36f70 | | ProoFVer | 79.25 | Not recorded | 68fa6cf81e209131cea36f70 | | DominikS | 79.16 | Not recorded | 68fa6cf81e209131cea36f70 | | ProoFVer-A | 77.16 | Not recorded | 68fa6cf81e209131cea36f70 | | DREAM | 76.85 | Not recorded | 68fa709a1e209131cea4c257 | | DOMLIN++ | 76.6 | Not recorded | 68fa709a1e209131cea4c257 | | CorefBERT | 75.96 | Not recorded | 68fa6cf81e209131cea36f70 | | ROBERTaBase Flexible (IE1) + MSPP (REk) | 75.36 | Not recorded | 68fa709a1e209131cea4c257 | | Joint Pre IEk + FT IE1 | 74.39 | Not recorded | 68fa6cf11e209131cea36af8 | | ROBERTaBase Fixed (IE1) + MSPP (IEk) | 74.39 | Not recorded | 68fa709a1e209131cea4c257 | | Joint Pre IEk + FT AE1 | 74.25 | Not recorded | 68fa6cf11e209131cea36af8 | | ProoFVer-MV | 74.18 | Not recorded | 68fa6cf81e209131cea36f70 | | Joint Pre AEk + FT AE1 | 74.13 | Not recorded | 68fa6cf11e209131cea36af8 | | ROBERTaBase Fixed (AE1) + MSPP (AEk) | 74.13 | Not recorded | 68fa709a1e209131cea4c257 | | KGAT | 74.07 | Not recorded | 68fa6cf11e209131cea36af8 | | KGAT | 74.07 | Not recorded | 68fa6cf81e209131cea36f70 | | KGAT with RoBERTaBase | 74.07 | Not recorded | 68fa709a1e209131cea4c257 | | Joint Pre AEk + FT IE1 | 73.77 | Not recorded | 68fa6cf11e209131cea36af8 | | Joint RoBERTa-Base + FT IE1 | 73.56 | Not recorded | 68fa6cf11e209131cea36af8 | | ROBERTaBase Flexible (IE1) | 73.56 | Not recorded | 68fa709a1e209131cea4c257 | | ROBERTaBase Fixed (IE1) | 73.01 | Not recorded | 68fa709a1e209131cea4c257 | | ROBERTaBase Fixed (AE1) | 72.94 | Not recorded | 68fa709a1e209131cea4c257 | | Transformer-XH | 72.39 | Not recorded | 68fa6cf11e209131cea36af8 | | Transformer-XH | 72.39 | Not recorded | 68fa709a1e209131cea4c257 | | GEAR | 71.6 | Not recorded | 68fa661b1e209131cea0e57c | | GEAR | 71.6 | Not recorded | 68fa6cf11e209131cea36af8 | | GEAR | 71.6 | Not recorded | 68fa709a1e209131cea4c257 | | BERT Concat | 71.01 | Not recorded | 68fa661b1e209131cea0e57c | | BERT Pair | 69.75 | Not recorded | 68fa661b1e209131cea0e57c | | UNC NLP | 68.21 | Not recorded | 68fa661b1e209131cea0e57c | | UCL MRG | 67.62 | Not recorded | 68fa661b1e209131cea0e57c | | Athene | 65.46 | Not recorded | 68fa661b1e209131cea0e57c | ### ProoFVer-SB LA Score: 79.47. Result ID: 698f6f633a6c314849f88c03. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retriever":"Stammbach (2021)"}. This record does not assert independent reproduction. ### ProoFVer LA Score: 79.25. Result ID: 698f6f633a6c314849f88bfe. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retriever":"Liu et al. (2020)"}. This record does not assert independent reproduction. ### DominikS LA Score: 79.16. Result ID: 698f6f633a6c314849f88c04. [Reporting paper]() Recorded evaluation configuration: {"Retriever":"Stammbach (2021)"}. This record does not assert independent reproduction. ### ProoFVer-A LA Score: 77.16. Result ID: 698f6f633a6c314849f88c00. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retriever":"Liu et al. (2020)"}. This record does not assert independent reproduction. ### DREAM LA Score: 76.85. Result ID: 69914fa13a6c314849073b40. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"XLNet-Large"}. This record does not assert independent reproduction. ### DOMLIN++ LA Score: 76.6. Result ID: 69914fa13a6c314849073b3f. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Large"}. This record does not assert independent reproduction. ### CorefBERT LA Score: 75.96. Result ID: 698f6f633a6c314849f88c02. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retriever":"Liu et al. (2020)"}. This record does not assert independent reproduction. ### ROBERTaBase Flexible (IE1) + MSPP (REk) LA Score: 75.36. Result ID: 69914fa13a6c314849073b46. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base","Sequence Type":"Flexible","Prediction Head":"IE1","MSPP Pre-training":"REk"}. This record does not assert independent reproduction. ### Joint Pre IEk + FT IE1 LA Score: 74.39. Result ID: 698f4cef3a6c314849f7c398. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-training head":"IEk","Fine-tuning head":"IE1"}. This record does not assert independent reproduction. ### ROBERTaBase Fixed (IE1) + MSPP (IEk) LA Score: 74.39. Result ID: 69914fa13a6c314849073b44. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base","Sequence Type":"Fixed","Prediction Head":"IE1","MSPP Pre-training":"IEk"}. This record does not assert independent reproduction. ### Joint Pre IEk + FT AE1 LA Score: 74.25. Result ID: 698f4cef3a6c314849f7c399. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-training head":"IEk","Fine-tuning head":"AE1"}. This record does not assert independent reproduction. ### ProoFVer-MV LA Score: 74.18. Result ID: 698f6f633a6c314849f88bff. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retriever":"Liu et al. (2020)"}. This record does not assert independent reproduction. ### Joint Pre AEk + FT AE1 LA Score: 74.13. Result ID: 698f4cef3a6c314849f7c39b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-training head":"AEk","Fine-tuning head":"AE1"}. This record does not assert independent reproduction. ### ROBERTaBase Fixed (AE1) + MSPP (AEk) LA Score: 74.13. Result ID: 69914fa13a6c314849073b45. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base","Sequence Type":"Fixed","Prediction Head":"AE1","MSPP Pre-training":"AEk"}. This record does not assert independent reproduction. ### KGAT LA Score: 74.07. Result ID: 698f4cef3a6c314849f7c395. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base"}. This record does not assert independent reproduction. ### KGAT LA Score: 74.07. Result ID: 698f6f633a6c314849f88c01. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retriever":"Liu et al. (2020)"}. This record does not assert independent reproduction. ### KGAT with RoBERTaBase LA Score: 74.07. Result ID: 69914fa13a6c314849073b3d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base"}. This record does not assert independent reproduction. ### Joint Pre AEk + FT IE1 LA Score: 73.77. Result ID: 698f4cef3a6c314849f7c39a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Pre-training head":"AEk","Fine-tuning head":"IE1"}. This record does not assert independent reproduction. ### Joint RoBERTa-Base + FT IE1 LA Score: 73.56. Result ID: 698f4cef3a6c314849f7c397. [Reporting paper]() Recorded evaluation configuration: {"Approach":"Joint","Backbone":"RoBERTa-Base","Fine-tuning head":"IE1"}. This record does not assert independent reproduction. ### ROBERTaBase Flexible (IE1) LA Score: 73.56. Result ID: 69914fa13a6c314849073b43. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base","Sequence Type":"Flexible","Prediction Head":"IE1","MSPP Pre-training":false}. This record does not assert independent reproduction. ### ROBERTaBase Fixed (IE1) LA Score: 73.01. Result ID: 69914fa13a6c314849073b41. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base","Sequence Type":"Fixed","Prediction Head":"IE1","MSPP Pre-training":false}. This record does not assert independent reproduction. ### ROBERTaBase Fixed (AE1) LA Score: 72.94. Result ID: 69914fa13a6c314849073b42. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"RoBERTa-Base","Sequence Type":"Fixed","Prediction Head":"AE1","MSPP Pre-training":false}. This record does not assert independent reproduction. ### Transformer-XH LA Score: 72.39. Result ID: 698f4cef3a6c314849f7c396. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Transformer-XH LA Score: 72.39. Result ID: 69914fa13a6c314849073b3e. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### GEAR LA Score: 71.6. Result ID: 698cdad78b517eb4331d3081. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### GEAR LA Score: 71.6. Result ID: 698f4cef3a6c314849f7c394. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### GEAR LA Score: 71.6. Result ID: 69914fa13a6c314849073b3c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### BERT Concat LA Score: 71.01. Result ID: 698cdad78b517eb4331d3080. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### BERT Pair LA Score: 69.75. Result ID: 698cdad78b517eb4331d307f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### UNC NLP LA Score: 68.21. Result ID: 698cdad78b517eb4331d307e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### UCL MRG LA Score: 67.62. Result ID: 698cdad78b517eb4331d307d. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Athene LA Score: 65.46. Result ID: 698cdad78b517eb4331d307c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.