# Multi-needle retrieval on MRCR-8needle 100-example

Source: [Multi-needle retrieval on MRCR-8needle 100-example](<https://www.sota2.com/research/sota/multi-needle-retrieval-on-mrcr-8needle-100-example>)

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

Data updated: 2026-06-04T08:49:50.854Z

Evaluation of model performance on the specific 100-example subset of the MRCR dataset configured for 8 needles, measured by Sequence match accuracy.

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 | Multi-needle retrieval |
| Dataset (including any stated split) | MRCR-8needle 100-example |
| Metric | Sequence Match Accuracy |
| Unit | PERCENT |
| 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

- [Sequence Match Accuracy](<https://www.sota2.com/research/sota/multi-needle-retrieval-on-mrcr-8needle-100-example.md?metric=Sequence%20Match%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Sequence Match Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| InfoMem | 27.9 | Not recorded | 6a210666c10484bac23eff3b |
| Qwen2.5-1.5B-Instruct | 26 | Not recorded | 6a210666c10484bac23eff3b |
| ReMemR1 | 9.2 | Not recorded | 6a210666c10484bac23eff3b |
| Outcome-only GRPO | 6.3 | Not recorded | 6a210666c10484bac23eff3b |

### InfoMem

Sequence Match Accuracy: 27.9. Result ID: 6a213c25fbc373592c1ce42f.

[Reporting paper](<https://www.sota2.com/research/paper/6a210666c10484bac23eff3b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a210666c10484bac23eff3b.md>)

Recorded evaluation configuration: {"base\_model":"Qwen2.5-1.5B-Instruct","beta":"0.2"}. This record does not assert independent reproduction.

### Qwen2.5-1.5B-Instruct

Sequence Match Accuracy: 26. Result ID: 6a213c25fbc373592c1ce42d.

[Reporting paper](<https://www.sota2.com/research/paper/6a210666c10484bac23eff3b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"status":"Base Model"}. This record does not assert independent reproduction.

### ReMemR1

Sequence Match Accuracy: 9.2. Result ID: 6a213c25fbc373592c1ce430.

[Reporting paper](<https://www.sota2.com/research/paper/6a210666c10484bac23eff3b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bb51e209131cea8a113.md>)

Recorded evaluation configuration: {"base\_model":"Qwen2.5-1.5B-Instruct","retrieval":"callback retrieval","framework":"chunk-wise memory-agent"}. This record does not assert independent reproduction.

### Outcome-only GRPO

Sequence Match Accuracy: 6.3. Result ID: 6a213c25fbc373592c1ce42e.

[Reporting paper](<https://www.sota2.com/research/paper/6a210666c10484bac23eff3b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79c71e209131cea7f569.md>)

Recorded evaluation configuration: {"base\_model":"Qwen2.5-1.5B-Instruct"}. This record does not assert independent reproduction.

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