# Time Series Question Answering on TimeSeriesExam synthetic (test)

Source: [Time Series Question Answering on TimeSeriesExam synthetic (test)](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test>)

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

Data updated: 2026-02-26T05:34:56.015Z

Evaluation of model performance on the Time Series Question Answering task using the synthetic test split of the TimeSeriesExam dataset, reporting component metrics such as Anomaly Detection, Causality, and Pattern Recognition.

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 | Time Series Question Answering |
| Dataset (including any stated split) | TimeSeriesExam synthetic (test) |
| Metric | Anomaly Detection |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [Anomaly Detection](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test.md?metric=Anomaly%20Detection>): Higher is better; unit: PERCENT.

- [Causality](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test.md?metric=Causality>): Higher is better; unit: PERCENT.

- [Noise Understanding](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test.md?metric=Noise%20Understanding>): Higher is better; unit: PERCENT.

- [Pattern Recognition](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test.md?metric=Pattern%20Recognition>): Higher is better; unit: PERCENT.

- [Similarity Analysis](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test.md?metric=Similarity%20Analysis>): Higher is better; unit: PERCENT.

- [Average Score](<https://www.sota2.com/research/sota/time-series-question-answering-on-timeseriesexam-synthetic-test.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Anomaly Detection | Result date | Report paper |
| --- | --- | --- | --- |
| TIMEOMNI-1 | 42.6 | Not recorded | 68fa7bc71e209131cea8a6ba |
| Qwen2.5-Instruct-7B | 37 | Not recorded | 68fa7bc71e209131cea8a6ba |
| Phi-3.5 | 28 | Not recorded | 68fa7bc71e209131cea8a6ba |

### TIMEOMNI-1

Anomaly Detection: 42.6. Result ID: 699ae5075698665ec4752cd0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6ba.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6ba.md>)

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

### Qwen2.5-Instruct-7B

Anomaly Detection: 37. Result ID: 699ae5075698665ec4752cce.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6ba.md>)

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

### Phi-3.5

Anomaly Detection: 28. Result ID: 699ae5075698665ec4752ccf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bc71e209131cea8a6ba.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75541e209131cea67378.md>)

Recorded evaluation configuration: {"Source":"Original Paper"}. This record does not assert independent reproduction.

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