# Personalized Retrieval on PersonaBench

Source: [Personalized Retrieval on PersonaBench](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench>)

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

Data updated: 2026-03-11T08:24:37.558Z

Evaluation of Personalized Retrieval performance on the PersonaBench dataset, measured using Recall@5 and Recall@10 across different information modalities (Time, Basic Info, Social Info, Preferences).

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 | Personalized Retrieval |
| Dataset (including any stated split) | PersonaBench |
| Metric | Latency@5 (Time) |
| Unit | TIME |
| Direction | Lower 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

- [Latency@5 (Time)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Latency%405%20(Time)>): Lower is better; unit: TIME.

- [Recall@5 (Basic Info)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%405%20(Basic%20Info)>): Higher is better; unit: PERCENT.

- [Recall@5 (Social Info)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%405%20(Social%20Info)>): Higher is better; unit: PERCENT.

- [Recall@5 (Pref Easy)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%405%20(Pref%20Easy)>): Higher is better; unit: PERCENT.

- [Recall@5 (Pref Hard)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%405%20(Pref%20Hard)>): Higher is better; unit: PERCENT.

- [Recall@5 (Overall)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%405%20(Overall)>): Higher is better; unit: PERCENT.

- [Latency@10 (Time)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Latency%4010%20(Time)>): Lower is better; unit: TIME.

- [Recall@10 (Basic Info)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%4010%20(Basic%20Info)>): Higher is better; unit: PERCENT.

- [Recall@10 (Social Info)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%4010%20(Social%20Info)>): Higher is better; unit: PERCENT.

- [Recall@10 (Pref Easy)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%4010%20(Pref%20Easy)>): Higher is better; unit: PERCENT.

- [Recall@10 (Pref Hard)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%4010%20(Pref%20Hard)>): Higher is better; unit: PERCENT.

- [Recall@10 (Overall)](<https://www.sota2.com/research/sota/personalized-retrieval-on-personabench.md?metric=Recall%4010%20(Overall)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Latency@5 (Time) | Result date | Report paper |
| --- | --- | --- | --- |
| Familiarity | 7.64 | Not recorded | 69b0f7232d603e6de3e45e87 |
| RF-Mem | 8.33 | Not recorded | 69b0f7232d603e6de3e45e87 |
| Familiarity | 8.4 | Not recorded | 69b0f7232d603e6de3e45e87 |
| Familiarity | 8.92 | Not recorded | 69b0f7232d603e6de3e45e87 |
| RF-Mem | 9.16 | Not recorded | 69b0f7232d603e6de3e45e87 |
| Recollection | 9.65 | Not recorded | 69b0f7232d603e6de3e45e87 |
| RF-Mem | 10.14 | Not recorded | 69b0f7232d603e6de3e45e87 |
| Recollection | 10.94 | Not recorded | 69b0f7232d603e6de3e45e87 |
| Recollection | 12.14 | Not recorded | 69b0f7232d603e6de3e45e87 |

### Familiarity

Latency@5 (Time): 7.64. Result ID: 69b126c16dbb7075b95b8242.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77481e209131cea7239c.md>)

Recorded evaluation configuration: {"Backbone":"all-mpnet-base-v2"}. This record does not assert independent reproduction.

### RF-Mem

Latency@5 (Time): 8.33. Result ID: 69b126c16dbb7075b95b8244.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

Recorded evaluation configuration: {"Backbone":"all-mpnet-base-v2"}. This record does not assert independent reproduction.

### Familiarity

Latency@5 (Time): 8.4. Result ID: 69b126c16dbb7075b95b823f.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77481e209131cea7239c.md>)

Recorded evaluation configuration: {"Backbone":"multi-qa-MiniLM-L6-cos-v1"}. This record does not assert independent reproduction.

### Familiarity

Latency@5 (Time): 8.92. Result ID: 69b126c16dbb7075b95b8245.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77481e209131cea7239c.md>)

Recorded evaluation configuration: {"Backbone":"BAAI/bge-base-en-v1.5"}. This record does not assert independent reproduction.

### RF-Mem

Latency@5 (Time): 9.16. Result ID: 69b126c16dbb7075b95b8241.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

Recorded evaluation configuration: {"Backbone":"multi-qa-MiniLM-L6-cos-v1"}. This record does not assert independent reproduction.

### Recollection

Latency@5 (Time): 9.65. Result ID: 69b126c16dbb7075b95b8240.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

Recorded evaluation configuration: {"Backbone":"multi-qa-MiniLM-L6-cos-v1"}. This record does not assert independent reproduction.

### RF-Mem

Latency@5 (Time): 10.14. Result ID: 69b126c16dbb7075b95b8247.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

Recorded evaluation configuration: {"Backbone":"BAAI/bge-base-en-v1.5"}. This record does not assert independent reproduction.

### Recollection

Latency@5 (Time): 10.94. Result ID: 69b126c16dbb7075b95b8243.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

Recorded evaluation configuration: {"Backbone":"all-mpnet-base-v2"}. This record does not assert independent reproduction.

### Recollection

Latency@5 (Time): 12.14. Result ID: 69b126c16dbb7075b95b8246.

[Reporting paper](<https://www.sota2.com/research/paper/69b0f7232d603e6de3e45e87.md>)

Recorded evaluation configuration: {"Backbone":"BAAI/bge-base-en-v1.5"}. This record does not assert independent reproduction.

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