Model Card: Appeal Success Prediction

Model Description

• Developed by: Bianka Šimková
• Model type: Causal LLM fine-tuned for sequence classification
• Language: Slovak
• License: Apache license 2.0
• Base model: slovak-nlp/mistral-sk-7b
• Fine-tuned for: Diploma thesis — Using large language models for analyzing decisions of Slovak courts

Intended Use

Primary Use

Out-of-Scope Use

• Court decisions in languages other than Slovak
• Legal advice or real-world legal decision support — this model is a research prototype and should not be used for any actual legal proceedings

Training Data

• Data source: Open data portal of the Slovak Republic
• Dataset format: JSON
• Preprocessing: Text cleaning, segmentation into logical parts, sampling of examples under 3000 tokens

Training Procedure

• Base model: slovak-nlp/mistral-sk-7b
• Fine-tuning method: LoRA
• Framework: Hugging Face Transformers
• Hardware: NVIDIA RTX A6000
• Epochs: 3
• Batch size: 1

Evaluation Results

Labels

Limitations

• Trained on Slovak court decisions only; performance on other legal systems or languages is not guaranteed
• The model reflects patterns in historical court data, which may contain biases
• Not intended for production legal use

Citation

@mastersthesis{Simkova2026,
  author    = {Bianka Šimková},
  title     = {Using large language models for analyzing decisions of Slovak courts},
  school    = {Institute of Artificial Intelligence,
Faculty of Electrical Engineering and Informatics, 
Technical University of Košice},
  year      = {2026}
}