mertkayacs/jevalt-data
JevAlt training data: text classification in English, Turkish and German Decision-training data for small language models: situations, questions and probability labels in English, Turkish and German. Used to train Deem-4B, Karar-4B and Wähler-4B. Karar-4B scores 96.8% against Kev-4B's 87.1% on Turkish decisions from the same data pipeline. The 96.8% score is Karar-4B's result. Quick facts Item Value Languages English, Turkish, German English rows 9… See the full description on the dataset page: https://huggingface.co/datasets/mertkayacs/jevalt-data.
| Author / organization | mertkayacs |
|---|---|
| License (as declared) | other |
| Downloads | 328 (as reported by Hugging Face) |
| Likes | 1 |
| Created | 1 October 2026 |
| Last updated at source | 15 October 2026 |
Popularity, last 3 days
328 now (+23 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- mertkayacs/Wahler-4BModel · used to train
- mertkayacs/Wahler-4B-GGUFModel · used to train
- mertkayacs/Karar-4BModel · used to train
- mertkayacs/Karar-4B-GGUFModel · used to train
- mradermacher/Karar-4B-GGUFModel · used to train
- mertkayacs/Deem-4BModel · used to train
- mradermacher/Wahler-4B-GGUFModel · used to train
- mradermacher/Deem-4B-GGUFModel · used to train
- mertkayacs/Deem-4B-GGUFModel · used to train
More from mertkayacs
- mertkayacs/jevalt-benchDataset
- mertkayacs/emberwick-videosDataset
- mertkayacs/tholos-benchDataset
- mertkayacs/tholos-trajectoriesDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.