# Reliability of post-edit LLMs on Wikipedia Source: [Reliability of post-edit LLMs on Wikipedia]() Collection index: [llms.txt]() Data updated: 2026-02-27T05:58:26.580Z Evaluation of LLM reliability following post-editing actions, using the Wikipedia corpus, measured by BLEU and METEOR 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 | Reliability of post-edit LLMs | | Dataset (including any stated split) | Wikipedia | | Metric | BLEU | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 36 of 36 | | 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 - [BLEU](): Higher is better; unit: PERCENT. - [METEOR](): Higher is better; unit: PERCENT. ## Recorded results | Method | BLEU | Result date | Report paper | | --- | --- | --- | --- | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | GRACE | 100 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 97.5 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 96.6 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 96.2 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 95.4 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 94.6 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 94.1 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 92.5 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 92.2 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 92 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 91.4 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 91.2 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 90.7 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 90.6 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 90.4 | Not recorded | 68fa7acf1e209131cea853d3 | | PME | 90.4 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 88.7 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 87.8 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 87.5 | Not recorded | 68fa7acf1e209131cea853d3 | | MEMIT | 86.9 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 86.6 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 83.7 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 83.5 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 83 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 82 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 74.9 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 73.4 | Not recorded | 68fa7acf1e209131cea853d3 | | DeMem | 73 | Not recorded | 68fa7acf1e209131cea853d3 | ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e87a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"email"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e87e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"phone"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e882. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"URL"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e886. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"email"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e88a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"phone"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e88e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"URL"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e892. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"email"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e896. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"phone"}. This record does not assert independent reproduction. ### GRACE BLEU: 100. Result ID: 69a1326e35dfcabb00e4e89a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"URL"}. This record does not assert independent reproduction. ### PME BLEU: 97.5. Result ID: 69a1326e35dfcabb00e4e880. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"URL"}. This record does not assert independent reproduction. ### PME BLEU: 96.6. Result ID: 69a1326e35dfcabb00e4e87c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"phone"}. This record does not assert independent reproduction. ### PME BLEU: 96.2. Result ID: 69a1326e35dfcabb00e4e894. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"phone"}. This record does not assert independent reproduction. ### PME BLEU: 95.4. Result ID: 69a1326e35dfcabb00e4e890. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"email"}. This record does not assert independent reproduction. ### PME BLEU: 94.6. Result ID: 69a1326e35dfcabb00e4e888. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"phone"}. This record does not assert independent reproduction. ### PME BLEU: 94.1. Result ID: 69a1326e35dfcabb00e4e878. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"email"}. This record does not assert independent reproduction. ### MEMIT BLEU: 92.5. Result ID: 69a1326e35dfcabb00e4e889. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"phone"}. This record does not assert independent reproduction. ### PME BLEU: 92.2. Result ID: 69a1326e35dfcabb00e4e884. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"email"}. This record does not assert independent reproduction. ### MEMIT BLEU: 92. Result ID: 69a1326e35dfcabb00e4e87d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"phone"}. This record does not assert independent reproduction. ### MEMIT BLEU: 91.4. Result ID: 69a1326e35dfcabb00e4e885. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"email"}. This record does not assert independent reproduction. ### PME BLEU: 91.2. Result ID: 69a1326e35dfcabb00e4e88c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"URL"}. This record does not assert independent reproduction. ### MEMIT BLEU: 90.7. Result ID: 69a1326e35dfcabb00e4e88d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"URL"}. This record does not assert independent reproduction. ### MEMIT BLEU: 90.6. Result ID: 69a1326e35dfcabb00e4e891. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"email"}. This record does not assert independent reproduction. ### MEMIT BLEU: 90.4. Result ID: 69a1326e35dfcabb00e4e879. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"email"}. This record does not assert independent reproduction. ### PME BLEU: 90.4. Result ID: 69a1326e35dfcabb00e4e898. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"URL"}. This record does not assert independent reproduction. ### MEMIT BLEU: 88.7. Result ID: 69a1326e35dfcabb00e4e881. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"URL"}. This record does not assert independent reproduction. ### MEMIT BLEU: 87.8. Result ID: 69a1326e35dfcabb00e4e899. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"URL"}. This record does not assert independent reproduction. ### DeMem BLEU: 87.5. Result ID: 69a1326e35dfcabb00e4e87b. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"email"}. This record does not assert independent reproduction. ### MEMIT BLEU: 86.9. Result ID: 69a1326e35dfcabb00e4e895. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"phone"}. This record does not assert independent reproduction. ### DeMem BLEU: 86.6. Result ID: 69a1326e35dfcabb00e4e883. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"URL"}. This record does not assert independent reproduction. ### DeMem BLEU: 83.7. Result ID: 69a1326e35dfcabb00e4e88f. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"URL"}. This record does not assert independent reproduction. ### DeMem BLEU: 83.5. Result ID: 69a1326e35dfcabb00e4e87f. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT Neo 1.3B","PII":"phone"}. This record does not assert independent reproduction. ### DeMem BLEU: 83. Result ID: 69a1326e35dfcabb00e4e887. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"email"}. This record does not assert independent reproduction. ### DeMem BLEU: 82. Result ID: 69a1326e35dfcabb00e4e88b. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT Neo 2.7B","PII":"phone"}. This record does not assert independent reproduction. ### DeMem BLEU: 74.9. Result ID: 69a1326e35dfcabb00e4e893. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"email"}. This record does not assert independent reproduction. ### DeMem BLEU: 73.4. Result ID: 69a1326e35dfcabb00e4e89b. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"URL"}. This record does not assert independent reproduction. ### DeMem BLEU: 73. Result ID: 69a1326e35dfcabb00e4e897. [Reporting paper]() Recorded evaluation configuration: {"Model":"GPT-J 6B","PII":"phone"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.