# End-of-Turn Detection on 357 conversation samples

Source: [End-of-Turn Detection on 357 conversation samples](<https://www.sota2.com/research/sota/end-of-turn-detection-on-357-conversation-samples>)

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

Data updated: 2026-03-17T07:50:21.228Z

Evaluation of a model's ability to accurately detect the end of a conversational turn using the 357 conversation samples dataset. Metrics include latency, precision, recall, F1 score, and 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 | End-of-Turn Detection |
| Dataset (including any stated split) | 357 conversation samples |
| Metric | Latency (ms) |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 2 of 2 |
| 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 (ms)](<https://www.sota2.com/research/sota/end-of-turn-detection-on-357-conversation-samples.md?metric=Latency%20(ms)>): Lower is better; unit: TIME.

- [Precision](<https://www.sota2.com/research/sota/end-of-turn-detection-on-357-conversation-samples.md?metric=Precision>): Higher is better; unit: PERCENT.

- [Recall](<https://www.sota2.com/research/sota/end-of-turn-detection-on-357-conversation-samples.md?metric=Recall>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/end-of-turn-detection-on-357-conversation-samples.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

- [Accuracy](<https://www.sota2.com/research/sota/end-of-turn-detection-on-357-conversation-samples.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Latency (ms) | Result date | Report paper |
| --- | --- | --- | --- |
| Hierarchical EOT | 36 | Not recorded | 69b8e0090636e5bf4afbe9ef |
| Smart Turn v3 | 900 | Not recorded | 69b8e0090636e5bf4afbe9ef |

### Hierarchical EOT

Latency (ms): 36. Result ID: 69b907b447f34ce924a14af5.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e0090636e5bf4afbe9ef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b8e0090636e5bf4afbe9ef.md>)

Recorded evaluation configuration: {"Params":"1.14M"}. This record does not assert independent reproduction.

### Smart Turn v3

Latency (ms): 900. Result ID: 69b907b447f34ce924a14af6.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e0090636e5bf4afbe9ef.md>)

Recorded evaluation configuration: {"Params":"8–12M"}. This record does not assert independent reproduction.

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