Bielik Guard: Efficient Polish Language Safety Classifiers for LLM Content Moderation

• URL: http://arxiv.org/abs/2602.07954v3
• Date: Fri, 13 Feb 2026 15:33:36 GMT
• Title: Bielik Guard: Efficient Polish Language Safety Classifiers for LLM Content Moderation
• Abstract summary: Bielik Guard is a family of compact Polish language safety classifiers.<n>Fine-tuned on a community-annotated dataset of 6,885 Polish texts, these models classify content across five safety categories: Hate/Aggression, Vulgarities, Sexual Content, Crime, and Self-Harm.
• Score: 0.4349640169711269
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As Large Language Models (LLMs) become increasingly deployed in Polish language applications, the need for efficient and accurate content safety classifiers has become paramount. We present Bielik Guard, a family of compact Polish language safety classifiers comprising two model variants: a 0.1B parameter model based on MMLW-RoBERTa-base and a 0.5B parameter model based on PKOBP/polish-roberta-8k. Fine-tuned on a community-annotated dataset of 6,885 Polish texts, these models classify content across five safety categories: Hate/Aggression, Vulgarities, Sexual Content, Crime, and Self-Harm. Our evaluation demonstrates that both models achieve strong performance on multiple benchmarks. The 0.5B variant offers the best overall discrimination capability with F1 scores of 0.791 (micro) and 0.785 (macro) on the test set, while the 0.1B variant demonstrates exceptional efficiency. Notably, Bielik Guard 0.1B v1.1 achieves superior precision (77.65%) and very low false positive rate (0.63%) on real user prompts, outperforming HerBERT-PL-Guard (31.55% precision, 4.70% FPR) despite identical model size. The models are publicly available and designed to provide appropriate responses rather than simple content blocking, particularly for sensitive categories like self-harm.

Related papers

• Models for minimalist RAG: B1ade 335M Embedding and 1B Parameter Small Language Models [2.1665689529884697]
  We present B1ade, an efficient RAG architecture comprising two purpose-built components: a compact embedding model and a purpose-built SLM.<n>B1ade-embed, a 335M parameter retrieval model constructed via parameter-free fusion of five pretrained encoders achieves top MTEB scores among sub-500M models with zero additional training.<n>B1ade-1B cites retrieved passages in 42.4% of responses, exceeding the attribution rate of its training distribution by 5.5 percentage points.
  arXiv  Detail & Related papers  (2026-07-29T22:41:25Z)
• Selecting Open-Weight Language Models for Zero-Shot Intent Classification: A Systematic Evaluation of 41 Models [0.0]
  We present a systematic zero-shot evaluation of 41 open-weight language models spanning 15 families and the 135M--9B parameter range.<n>We analyze confidence calibration, robustness to realistic input perturbations, statistical reliability of model rankings, deployment efficiency, and benchmark saturation.<n>These findings provide practical guidance for selecting and evaluating open-weight language models for intent classification.
  arXiv  Detail & Related papers  (2026-07-29T19:42:33Z)
• Sub-Billion, Super-Frontier: Small Language Models Rival Zero-Shot Frontier LLMs on General and Literary Relation Extraction [5.070542698701158]
  Large language models (LLMs) achieve strong relation extraction (RE)<n>We investigate how far small language models (SLMs) can close this gap across general-domain and literary text.
  arXiv  Detail & Related papers  (2026-06-21T17:24:31Z)
• Models Can Model, But Can't Bind: Structured Grounding in Text-to-Optimization [54.749573452394664]
  We find that accuracy collapses as instance data grows, even when the formulation itself is simple.<n>We address this via a simple inference-time approach, BIND, which externalizes numeric data to structured files so the model binds data rather than from prompt prompt.<n>We validate our hypothesis by finetuning a model exclusively on binding and show that it outperforms end-to-end SFT and RL across three structurally distinct optimization categories.
  arXiv  Detail & Related papers  (2026-05-20T21:25:41Z)
• Benchmarking Cross-Lingual Semantic Alignment in Multilingual Embeddings [0.0]
  Task-driven benchmarks (MTEB) may mask fundamental alignment shortcomings.<n>We introduce Semantic Affinity (SA), a bounded (between 0 and 1) metric measuring inter-lingual to intra-lingual spread ratio.<n> Benchmarking 13 models across 4 datasets (52 experiments) reveals a three-tier structure.
  arXiv  Detail & Related papers  (2025-12-29T14:32:57Z)
• XplaiNLP at CheckThat! 2025: Multilingual Subjectivity Detection with Finetuned Transformers and Prompt-Based Inference with Large Language Models [2.749729059235755]
  This notebook reports the Xplai submission to the CheckThat! 2025 shared task on multilingual subjectivity detection.<n>We evaluate two approaches: supervised fine-tuning of transformer encoders, EuroBERT, XLM-RoBERTa, and German-BERT, on monolingual and machine-translated training data.<n>For German, a German-BERT model fine-tuned on translated training data from typologically related languages yields competitive performance over the baseline.
  arXiv  Detail & Related papers  (2025-09-15T16:53:41Z)
• Mario at EXIST 2025: A Simple Gateway to Effective Multilingual Sexism Detection [8.40042895828361]
  EXIST 2025 Task 1 addresses text-based sexism detection in English and Spanish tweets through hierarchical Low-Rank Adaptation (LoRA) of Llama 3.1 8B.<n>Our method introduces conditional adapter routing that explicitly models dependencies across three hierarchically structured subtasks.<n>Our approach reduces training time by 75% and model storage by 98%, while achieving competitive performance across all subtasks.
  arXiv  Detail & Related papers  (2025-07-15T05:30:32Z)
• PL-Guard: Benchmarking Language Model Safety for Polish [43.39208658482427]
  We introduce a manually annotated benchmark dataset for language model safety classification in Polish.<n>We also create adversarially perturbed variants of these samples designed to challenge model robustness.<n>We train these models using different combinations of annotated data and evaluate their performance, comparing it against publicly available guard models.
  arXiv  Detail & Related papers  (2025-06-19T13:56:41Z)
• Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval [49.1574468325115]
  We introduce Amharic-specific dense retrieval models based on pre-trained Amharic BERT and RoBERTa backbones.<n>Our proposed RoBERTa-Base-Amharic-Embed model (110M parameters) achieves a 17.6% relative improvement in MRR@10.<n>More compact variants, such as RoBERTa-Medium-Amharic-Embed (42M) remain competitive while being over 13x smaller.
  arXiv  Detail & Related papers  (2025-05-25T23:06:20Z)
• WorldPM: Scaling Human Preference Modeling [130.23230492612214]
  We propose World Preference Modeling$ (WorldPM) to emphasize this scaling potential.<n>We collect preference data from public forums covering diverse user communities.<n>We conduct extensive training using 15M-scale data across models ranging from 1.5B to 72B parameters.
  arXiv  Detail & Related papers  (2025-05-15T17:38:37Z)
• Secret Breach Detection in Source Code with Large Language Models [2.5484785866796833]
  Leaking sensitive information in source code remains a persistent security threat.<n>This work aims to enhance secret detection in source code using large language models (LLMs)<n>We evaluate the feasibility of using fine-tuned, smaller models for local deployment.
  arXiv  Detail & Related papers  (2025-04-26T03:33:14Z)
• DataComp-LM: In search of the next generation of training sets for language models [200.5293181577585]
  DataComp for Language Models (DCLM) is a testbed for controlled dataset experiments with the goal of improving language models.<n>We provide a standardized corpus of 240T tokens extracted from Common Crawl, effective pretraining recipes based on the OpenLM framework, and a broad suite of 53 downstream evaluations.<n>Participants in the DCLM benchmark can experiment with data curation strategies such as deduplication, filtering, and data mixing at model scales ranging from 412M to 7B parameters.
  arXiv  Detail & Related papers  (2024-06-17T17:42:57Z)
• OneAligner: Zero-shot Cross-lingual Transfer with One Rich-Resource Language Pair for Low-Resource Sentence Retrieval [91.76575626229824]
  We present OneAligner, an alignment model specially designed for sentence retrieval tasks. When trained with all language pairs of a large-scale parallel multilingual corpus (OPUS-100), this model achieves the state-of-the-art result. We conclude through empirical results and analyses that the performance of the sentence alignment task depends mostly on the monolingual and parallel data size.
  arXiv  Detail & Related papers  (2022-05-17T19:52:42Z)
• DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing [117.41016786835452]
  This paper presents a new pre-trained language model, DeBERTaV3, which improves the original DeBERTa model. vanilla embedding sharing in ELECTRA hurts training efficiency and model performance. We propose a new gradient-disentangled embedding sharing method that avoids the tug-of-war dynamics.
  arXiv  Detail & Related papers  (2021-11-18T06:48:00Z)
• DeBERTa: Decoding-enhanced BERT with Disentangled Attention [119.77305080520718]
  We propose a new model architecture DeBERTa that improves the BERT and RoBERTa models using two novel techniques. We show that these techniques significantly improve the efficiency of model pre-training and the performance of both natural language understanding (NLU) and natural langauge generation (NLG) downstream tasks.
  arXiv  Detail & Related papers  (2020-06-05T19:54:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.