5 papers
Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content
Ihor Stepanov, Aleksandr Smechov
Real-time safety filtering for large language model (LLM) applications requires classifiers that can detect unsafe prompts, toxic language, jailbreak attempts, and unsafe responses…
GLiNER-Relex: A Unified Framework for Joint Named Entity Recognition and Relation Extraction
Ihor Stepanov, Oleksandr Lukashov, Mykhailo Shtopko +1
Joint named entity recognition (NER) and relation extraction (RE) is a fundamental task in natural language processing for constructing knowledge graphs from unstructured text. Whi…
The Million-Label NER: Breaking Scale Barriers with GLiNER bi-encoder
Ihor Stepanov, Mykhailo Shtopko, Dmytro Vodianytskyi +1
This paper introduces GLiNER-bi-Encoder, a novel architecture for Named Entity Recognition (NER) that harmonizes zero-shot flexibility with industrial-scale efficiency. While the o…
GLiClass: Generalist Lightweight Model for Sequence Classification Tasks
Ihor Stepanov, Mykhailo Shtopko, Dmytro Vodianytskyi +3
Classification is one of the most widespread tasks in AI applications, serving often as the first step in filtering, sorting, and categorizing data. Since modern AI systems must ha…
GLiNER-BioMed: A Suite of Efficient Models for Open Biomedical Named Entity Recognition
Anthony Yazdani, Ihor Stepanov, Douglas Teodoro
Biomedical named entity recognition (NER) presents unique challenges due to specialized vocabularies, the sheer volume of entities, and the continuous emergence of novel entities.…