# Lifelong Knowledge Editing on GPT-J 1024 Edits (test)

Source: [Lifelong Knowledge Editing on GPT-J 1024 Edits (test)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:04:06.202Z

Evaluation of Lifelong Knowledge Editing performance on the GPT-J model subjected to 1024 edits, measured using metrics such as Score (S), Effectiveness (ES), Generality (PS), Locality (NS), and Retention (ERS/ORS).

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 | Lifelong Knowledge Editing |
| Dataset (including any stated split) | Lifelong Editing on GPT-J 1024 edits (test) |
| Metric | Score (S) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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

- [Score (S)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test.md?metric=Score%20(S)>): Higher is better; unit: SCALAR.

- [Effectiveness (ES)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test.md?metric=Effectiveness%20(ES)>): Higher is better; unit: SCALAR.

- [Generality (PS)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test.md?metric=Generality%20(PS)>): Higher is better; unit: SCALAR.

- [Locality (NS)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test.md?metric=Locality%20(NS)>): Higher is better; unit: SCALAR.

- [Retention (ERS)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test.md?metric=Retention%20(ERS)>): Higher is better; unit: SCALAR.

- [Retention (ORS)](<https://www.sota2.com/research/sota/lifelong-knowledge-editing-on-gpt-j-1024-edits-test.md?metric=Retention%20(ORS)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Score (S) | Result date | Report paper |
| --- | --- | --- | --- |
| WilKE | 14.6 | Not recorded | 68fa73701e209131cea5c802 |
| ROME | 8.7 | Not recorded | 68fa73701e209131cea5c802 |
| MEND | 3.7 | Not recorded | 68fa73701e209131cea5c802 |
| MEMIT | 0 | Not recorded | 68fa73701e209131cea5c802 |

### WilKE

Score (S): 14.6. Result ID: 6992eb3b3a6c314849125b81.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73701e209131cea5c802.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73701e209131cea5c802.md>)

[Implementation](<https://github.com/ChenhuiHu/WilKE>)

Recorded evaluation configuration: {"Backbone":"GPT-J (6B)"}. This record does not assert independent reproduction.

### ROME

Score (S): 8.7. Result ID: 6992eb3b3a6c314849125b7f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73701e209131cea5c802.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e591e209131cea3f3c9.md>)

[Implementation](<https://rome.baulab.info/>)

Recorded evaluation configuration: {"Backbone":"GPT-J (6B)"}. This record does not assert independent reproduction.

### MEND

Score (S): 3.7. Result ID: 6992eb3b3a6c314849125b7e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73701e209131cea5c802.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cc61e209131cea35ea9.md>)

Recorded evaluation configuration: {"Backbone":"GPT-J (6B)"}. This record does not assert independent reproduction.

### MEMIT

Score (S): 0. Result ID: 6992eb3b3a6c314849125b80.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73701e209131cea5c802.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa703d1e209131cea49dc7.md>)

Recorded evaluation configuration: {"Backbone":"GPT-J (6B)"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.