# Generative Retrieval on Yambda (block 10)

Source: [Generative Retrieval on Yambda (block 10)](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10>)

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

Data updated: 2026-04-16T07:14:06.685Z

Evaluation of Generative Retrieval performance on the Yambda (block 10) dataset, measured using Recall@K and nDCG@K metrics.

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 | Generative Retrieval |
| Dataset (including any stated split) | Yambda (block 10) |
| Metric | Recall@10 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

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

- [Recall@100](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10.md?metric=Recall%40100>): Higher is better; unit: PERCENT.

- [Recall@500](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10.md?metric=Recall%40500>): Higher is better; unit: PERCENT.

- [nDCG@10](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10.md?metric=nDCG%4010>): Higher is better; unit: PERCENT.

- [nDCG@100](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10.md?metric=nDCG%40100>): Higher is better; unit: PERCENT.

- [nDCG@500](<https://www.sota2.com/research/sota/generative-retrieval-on-yambda-block-10.md?metric=nDCG%40500>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Recall@10 | Result date | Report paper |
| --- | --- | --- | --- |
| LETTER (FT-ours, Hungarian) | 1.44 | Not recorded | 69e06d177d0172140ec7ca6c |
| LETTER (Full) | 1.42 | Not recorded | 69e06d177d0172140ec7ca6c |
| LETTER (FT-old) | 1.41 | Not recorded | 69e06d177d0172140ec7ca6c |
| LETTER (FT-ours, Greedy) | 1.34 | Not recorded | 69e06d177d0172140ec7ca6c |
| LETTER (FT-new) | 1.31 | Not recorded | 69e06d177d0172140ec7ca6c |
| TIGER (FT) | 1.28 | Not recorded | 69e06d177d0172140ec7ca6c |
| TIGER (Base) | 1.12 | Not recorded | 69e06d177d0172140ec7ca6c |
| LETTER (Base) | 1.1 | Not recorded | 69e06d177d0172140ec7ca6c |

### LETTER (FT-ours, Hungarian)

Recall@10: 1.44. Result ID: 69e08c2347f34ce924c9605f.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

Recorded evaluation configuration: {"SID update policy":"Hungarian alignment"}. This record does not assert independent reproduction.

### LETTER (Full)

Recall@10: 1.42. Result ID: 69e08c2347f34ce924c96060.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c3c1e209131cea8d152.md>)

Recorded evaluation configuration: {"mode":"Full retraining"}. This record does not assert independent reproduction.

### LETTER (FT-old)

Recall@10: 1.41. Result ID: 69e08c2347f34ce924c9605c.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c3c1e209131cea8d152.md>)

Recorded evaluation configuration: {"SID update policy":"stale SIDs"}. This record does not assert independent reproduction.

### LETTER (FT-ours, Greedy)

Recall@10: 1.34. Result ID: 69e08c2347f34ce924c9605e.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

Recorded evaluation configuration: {"SID update policy":"Greedy alignment"}. This record does not assert independent reproduction.

### LETTER (FT-new)

Recall@10: 1.31. Result ID: 69e08c2347f34ce924c9605d.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c3c1e209131cea8d152.md>)

Recorded evaluation configuration: {"SID update policy":"fresh SIDs"}. This record does not assert independent reproduction.

### TIGER (FT)

Recall@10: 1.28. Result ID: 69e08c2347f34ce924c9605a.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa710f1e209131cea4ef39.md>)

Recorded evaluation configuration: {"mode":"Fine-tuned"}. This record does not assert independent reproduction.

### TIGER (Base)

Recall@10: 1.12. Result ID: 69e08c2347f34ce924c96059.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa710f1e209131cea4ef39.md>)

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

### LETTER (Base)

Recall@10: 1.1. Result ID: 69e08c2347f34ce924c9605b.

[Reporting paper](<https://www.sota2.com/research/paper/69e06d177d0172140ec7ca6c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c3c1e209131cea8d152.md>)

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

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