All activity
TUNiBrain is a hybrid Korean morphological analyzer that combines deep learning and dictionary-based NLP to understand meaning, not just tokens. It detects proper nouns, disambiguates words like “눈” (eye vs snow), and outputs definitions, senses, and translations for every token. Built with CRF, Bi-LSTM, and semantic dictionaries, it powers more accurate Korean chatbots, search, and LLM pipelines.
TUNiBrain–Korean Morphological AnalyzerThe next-gen Korean morphological analyzer
조영택(yt)left a comment
Hey Product Hunters 👋 If you’ve ever built NLP systems for Korean, you already know the ugly truth: **tokenization alone is not understanding.** Most analyzers can split text — but they can’t tell whether **“눈” means an eye or snow**, **“배” is a fruit, a ship, or a stomach**, or whether a word is a **person, place, or brand**. So we built **TUNiBrain**. TUNiBrain is not just a tokenizer — it’s...
TUNiBrain–Korean Morphological AnalyzerThe next-gen Korean morphological analyzer
As AI evolves, so do threats. 😇Angel blocks prompt attacks like jailbreaking, leakage, and unethical content. 🛡️ Detect and stop threats before they harm your AI. 🔒Secure your AI with Angel and stay ahead! ⚔️
Angel - Prompt Attack Detection ModelKeep Your AI Safe from Malicious Prompt Attacks!
조영택(yt)left a comment
The Angel model demo lets you experience its security features in action through a turn-based interaction. 🔄 As you take turns with the AI—providing prompts and receiving responses—Angel works behind the scenes to detect prompt attacks in real time. 👀 If it spots any signs of a threat, like jailbreaking attempts or prompt leakage, you'll get an instant notification! 🚨 This shows how Angel keeps...
Angel - Prompt Attack Detection ModelKeep Your AI Safe from Malicious Prompt Attacks!
조영택(yt)left a comment
✨ **Meet Angel: The Ultimate Prompt Attack Detection Model!** 🛡️ In a world where AI models and services face increasingly sophisticated threats, **Angel** is here to keep your systems safe. 🚀 Here’s how **Angel** protects your AI from prompt-based attacks, categorized by the following types of vulnerabilities: - **🛑 Prompt Leaks:** Angel detects and prevents prompt leakage that could expose...
Angel - Prompt Attack Detection ModelKeep Your AI Safe from Malicious Prompt Attacks!