# Knowledge Retention on Sequential Unlearning 16-task Retained Data Avg Source: [Knowledge Retention on Sequential Unlearning 16-task Retained Data Avg]() Collection index: [llms.txt]() Data updated: 2026-03-24T15:36:24.694Z Evaluation of model knowledge retention capabilities across 16 tasks derived from sequential unlearning experiments, measured using metrics such as Specificity, Answer Rate, BERT Score, CLIP Score, and ROUGE-L. 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 | Knowledge Retention | | Dataset (including any stated split) | Sequential Unlearning 16-task Retained Data Avg | | Metric | Specificity | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 9 of 9 | | 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 - [Specificity](): Higher is better; unit: PERCENT. - [Answer Rate](): Higher is better; unit: PERCENT. - [BERT Score](): Higher is better; unit: PERCENT. - [CLIP Score](): Higher is better; unit: PERCENT. - [ROUGE-L Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Specificity | Result date | Report paper | | --- | --- | --- | --- | | Vicuna | 100 | Not recorded | 69c21a9e0636e5bf4afc4999 | | CORE | 95.09 | Not recorded | 69c21a9e0636e5bf4afc4999 | | O3 | 91.85 | Not recorded | 69c21a9e0636e5bf4afc4999 | | MoEAdapter | 90.48 | Not recorded | 69c21a9e0636e5bf4afc4999 | | EWC | 80.37 | Not recorded | 69c21a9e0636e5bf4afc4999 | | EProj | 77.65 | Not recorded | 69c21a9e0636e5bf4afc4999 | | LwF | 77.41 | Not recorded | 69c21a9e0636e5bf4afc4999 | | SCRUB | 72.5 | Not recorded | 69c21a9e0636e5bf4afc4999 | | GMM | 72.22 | Not recorded | 69c21a9e0636e5bf4afc4999 | ### Vicuna Specificity: 100. Result ID: 69c2af6647f34ce924abcbb1. [Reporting paper]() Recorded evaluation configuration: {"Mode":"Zero-shot"}. This record does not assert independent reproduction. ### CORE Specificity: 95.09. Result ID: 69c2af6647f34ce924abcbb9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### O3 Specificity: 91.85. Result ID: 69c2af6647f34ce924abcbb8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### MoEAdapter Specificity: 90.48. Result ID: 69c2af6647f34ce924abcbb7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### EWC Specificity: 80.37. Result ID: 69c2af6647f34ce924abcbb2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### EProj Specificity: 77.65. Result ID: 69c2af6647f34ce924abcbb5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LwF Specificity: 77.41. Result ID: 69c2af6647f34ce924abcbb3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### SCRUB Specificity: 72.5. Result ID: 69c2af6647f34ce924abcbb6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### GMM Specificity: 72.22. Result ID: 69c2af6647f34ce924abcbb4. [Reporting paper]() [Method origin paper]() [Implementation]() 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.