Dataset · AlicanKiraz0
Turkish SFT dataset with long responses
AlicanKiraz0/Turkish-SFT-Dataset-v1.0
Around 5,500 Turkish dialogue examples with long outputs for SFT.
What's inside
Contains around 5,500 rows in Turkish (over 20 million tokens, with 3,000 to 4,500 tokens per row). The data is provided in JSONL format with system, user, and assistant columns under the MIT license. Examples cover instruction following, reasoning, safe refusals, long context, and tool use.
What you can do with it
- SFT fine-tuning of language models in Turkish
- Training models for long response generation and reasoning
- Configuring safe refusals and tool-calling skills
- Tasks
- Text classification, Question answering, Text generation
- Languages
- Turkish
- Size
- 1K–10K rows
- Downloads this month
- 466
- Likes
- 53
Models trained on it
- Tuguberk/Qwen3.5-2B-Turkish-SFTText generation 582
- ufakai/ufakzeka-1Text generation 294
- sadecebirisii/Llama-3.1-8B-Turkish-Siber-Muhafiz 42
- sinanelms/Qwen2.5-7B-Instruct-Turkish-Legal-DoRAText generation 13
- Tuguberk/Qwen3.5-2B-Turkish-SFT-LoRAText generation 0
- aapo33/Cores-AI-TR-v1.0.0-85M-InstructText generation 0
How to load it
from datasets import load_dataset
ds = load_dataset("AlicanKiraz0/Turkish-SFT-Dataset-v1.0")