# LLM Unlearning on TOFU (Metrics: Agg., Mem., Priv., Util.) Source: [LLM Unlearning on TOFU (Metrics: Agg., Mem., Priv., Util.)]() Collection index: [llms.txt]() Data updated: 2026-05-08T09:17:39.399Z Evaluation of LLM Unlearning effectiveness on the TOFU dataset, reporting on Aggregated, Memory, Privacy, and Utility 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 | LLM Unlearning | | Dataset (including any stated split) | TOFU | | Metric | Aggregated Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 13 of 13 | | 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 - [Aggregated Score](): Higher is better; unit: PERCENT. - [Memory Score](): Higher is better; unit: PERCENT. - [Privacy Score](): Higher is better; unit: PERCENT. - [Utility Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Aggregated Score | Result date | Report paper | | --- | --- | --- | --- | | Retain | 61 | Not recorded | 69fd6e435497b3b670c895f9 | | ROSU | 58 | Not recorded | 69fd6e435497b3b670c895f9 | | RMU | 54 | Not recorded | 69fd6e435497b3b670c895f9 | | UAM | 50 | Not recorded | 69fd6e435497b3b670c895f9 | | SimNPO | 49 | Not recorded | 69fd6e435497b3b670c895f9 | | UNDIAL | 44 | Not recorded | 69fd6e435497b3b670c895f9 | | IdkNLL | 13 | Not recorded | 69fd6e435497b3b670c895f9 | | PCGrad | 6 | Not recorded | 69fd6e435497b3b670c895f9 | | AltPO | 5 | Not recorded | 69fd6e435497b3b670c895f9 | | IdkDPO | 3 | Not recorded | 69fd6e435497b3b670c895f9 | | NPO | 2 | Not recorded | 69fd6e435497b3b670c895f9 | | Init. ft. | 0 | Not recorded | 69fd6e435497b3b670c895f9 | | GradDiff | 0 | Not recorded | 69fd6e435497b3b670c895f9 | ### Retain Aggregated Score: 61. Result ID: 69fda93147f34ce924e0e0d1. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### ROSU Aggregated Score: 58. Result ID: 69fda93147f34ce924e0e0dc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### RMU Aggregated Score: 54. Result ID: 69fda93147f34ce924e0e0db. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### UAM Aggregated Score: 50. Result ID: 69fda93147f34ce924e0e0da. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### SimNPO Aggregated Score: 49. Result ID: 69fda93147f34ce924e0e0d9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### UNDIAL Aggregated Score: 44. Result ID: 69fda93147f34ce924e0e0d8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### IdkNLL Aggregated Score: 13. Result ID: 69fda93147f34ce924e0e0d7. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### PCGrad Aggregated Score: 6. Result ID: 69fda93147f34ce924e0e0d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### AltPO Aggregated Score: 5. Result ID: 69fda93147f34ce924e0e0d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### IdkDPO Aggregated Score: 3. Result ID: 69fda93147f34ce924e0e0d4. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### NPO Aggregated Score: 2. Result ID: 69fda93147f34ce924e0e0d3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### Init. ft. Aggregated Score: 0. Result ID: 69fda93147f34ce924e0e0d0. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. ### GradDiff Aggregated Score: 0. Result ID: 69fda93147f34ce924e0e0d2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.2-1B-Instruct"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.