# Stream Generation on StreamChat sequence length (4096×3, 4096×4\]

Source: [Stream Generation on StreamChat sequence length (4096×3, 4096×4\]](<https://www.sota2.com/research/sota/stream-generation-on-streamchat-sequence-length-4096-3-4096-4>)

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

Data updated: 2026-02-27T09:24:20.328Z

Evaluation of stream generation performance using the StreamChat dataset variant specified by sequence length constraints (4096x3 to 4096x4), measured using BLEU, ROUGE, and BERT-F 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 | Stream Generation |
| Dataset (including any stated split) | StreamChat sequence length (4096×3, 4096×4\] |
| Metric | BLEU |
| Unit | SCALAR |
| 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

- [BLEU](<https://www.sota2.com/research/sota/stream-generation-on-streamchat-sequence-length-4096-3-4096-4.md?metric=BLEU>): Higher is better; unit: SCALAR.

- [ROUGE](<https://www.sota2.com/research/sota/stream-generation-on-streamchat-sequence-length-4096-3-4096-4.md?metric=ROUGE>): Higher is better; unit: SCALAR.

- [BERT-F](<https://www.sota2.com/research/sota/stream-generation-on-streamchat-sequence-length-4096-3-4096-4.md?metric=BERT-F>): Higher is better; unit: SCALAR.

## Recorded results

| Method | BLEU | Result date | Report paper |
| --- | --- | --- | --- |
| ADORE | 36.7 | Not recorded | 68fa73e61e209131cea5ec78 |
| H2O(Rebuilt) | 30.4 | Not recorded | 68fa73e61e209131cea5ec78 |
| H2O | 27.3 | Not recorded | 68fa73e61e209131cea5ec78 |
| StreamingLLM | 21.7 | Not recorded | 68fa73e61e209131cea5ec78 |
| Window Attention | 19.9 | Not recorded | 68fa73e61e209131cea5ec78 |
| KV Compression | 16.9 | Not recorded | 68fa73e61e209131cea5ec78 |
| Full Attention | 2 | Not recorded | 68fa73e61e209131cea5ec78 |
| Strided Attention | 2 | Not recorded | 68fa73e61e209131cea5ec78 |

### ADORE

BLEU: 36.7. Result ID: 69a162b935dfcabb00e6b6c6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Implementation](<https://github.com/WHUIR/ADORE>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### H2O(Rebuilt)

BLEU: 30.4. Result ID: 69a162b935dfcabb00e6b6c5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### H2O

BLEU: 27.3. Result ID: 69a162b935dfcabb00e6b6c4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa718c1e209131cea51ead.md>)

[Implementation](<https://github.com/FMInference/H2O>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### StreamingLLM

BLEU: 21.7. Result ID: 69a162b935dfcabb00e6b6c3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72d71e209131cea58d79.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Window Attention

BLEU: 19.9. Result ID: 69a162b935dfcabb00e6b6c1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f501e209131cea44f4a.md>)

[Implementation](<https://github.com/SHI-Labs/Neighborhood-Attention-Transformer>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### KV Compression

BLEU: 16.9. Result ID: 69a162b935dfcabb00e6b6c2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ea1e209131cea4df67.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Full Attention

BLEU: 2. Result ID: 69a162b935dfcabb00e6b6bf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Strided Attention

BLEU: 2. Result ID: 69a162b935dfcabb00e6b6c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73e61e209131cea5ec78.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65aa1e209131cea0bc26.md>)

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.