license: apache-2.0
language: [tr]
task_categories: [text-classification]
tags: [turkish, llm-security, refusal, safety, evaluation, ai-safety, red-teaming, guardrails]
pretty_name: "Turkish LLM Refusal Markers"
size_categories: [n<1K]
configs: [{config_name: default, data_files: ref.jsonl}]
Turkish LLM Refusal Markers
from datasets import load_dataset
ds = load_dataset("fevziegeyurtsevenler/llm-refusal-markers-tr")
18 Turkish refusal markers (strong vs weak) for detecting genuine refusals when evaluating Turkish LLM safety / red-team results. Strong markers = explicit negation; weak = filler that alone doesn't prove refusal.
Useful to avoid mislabeling: a response echoing an attack word but containing a strong refusal marker is a defended case, not a leak. marker, type, language, gloss_en.
Schema
| column | meaning |
|---|---|
| marker | Turkish phrase/pattern |
| type | strong / weak |
| gloss_en | English gloss |
Related AltaySec resources
- 🕵️ uncloak — hidden prompt-injection scanner: GitHub · live demo
- 📚 Datasets: detection-rules · glossary · skills-in-the-wild · agent-skill-vulnerabilities
- 🌐 AltaySec · Açık Kaynak Lab
Citation
@misc{yurtsevenler2026llmrefusalmarkerstr,
title = {Turkish LLM Refusal Markers},
author = {Yurtsevenler, Fevzi Ege},
year = {2026},
publisher = {AltaySec / Hugging Face},
howpublished = {\url{https://huggingface.co/datasets/fevziegeyurtsevenler/llm-refusal-markers-tr}}
}
⚠️ Defensive / authorized-use only. Multilingual-first LLM/agent security research by AltaySec.