mlsanityData to 14 Oct 2026

GLiNER Guard: Unified Encoder Family for Production LLM Safety and Privacy

arXiv 2605.05277 ↗First posted 6 May 2026cs.CRcited by 2

On the map: LLM safety and security → Safety alignment, refusal and guardrails

Abstract (arXiv)

Production LLM systems require both safety moderation and PII detection under strict latency and cost constraints. This creates a trade-off: autoregressive moderators are accurate but expensive, while lightweight encoders are faster but less capable. We present GLiNER Guard (GLiGuard), a unified encoder that performs safety classification and PII detection in a single forward pass, simplifying safety pipelines. We introduce three variants: compact uni- and bi-encoders (145-147M) for high-throughput serving, and GLiGuard Omni (209M) for stronger moderation quality. Under dynamic batching on a single A100, the compact model reaches 193 requests/sec with P99 latency below 1s, achieving 1.6x higher throughput than GLiNER2. Omni remains competitive with much larger moderators on public safety benchmarks. We also release PII-Bench, a span-level benchmark for evaluating PII detection in end-to-end pipelines. Overall, encoder-based guardrails offer a practical low-cost alternative for always-on moderation. Models and benchmarks are released on HuggingFace.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security. Builds on 17 of its 22 references that are in the map. First posted on arXiv 6 May 2026.

Builds on

17 of 22 references.

Closest ancestors · 3

Same direction, newest first

  1. 2026YuFeng-XGuard: A Reasoning-Centric, Interpretable, and Flexible Guardrail Model for Large Language Models22 Jan
  2. 2025Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails15 Jan
  3. 2024WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs26 Jun

Other references · 14

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2020Longformer: The Long-Document Transformercited by 3,413
  3. 2025gpt-oss-120b & gpt-oss-20b Model Cardcited by 1,386
  4. 2021DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharingcited by 1,254
  5. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  6. 2024Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inferencecited by 554
  7. 2024A StrongREJECT for Empty Jailbreakscited by 385
  8. 2020CrossNER: Evaluating Cross-Domain Named Entity Recognitioncited by 104
  9. 2023GLiNER: Generalist Model for Named Entity Recognition using Bidirectional Transformercited by 88
  10. 2025PolyGuard: A Multilingual Safety Moderation Tool for 17 Languagescited by 59
  11. 2025mmBERT: A Modern Multilingual Encoder with Annealed Language Learningcited by 45
  12. 2025GLiClass: Generalist Lightweight Model for Sequence Classification Taskscited by 15
2 more
  1. 2025GLiNER2: An Efficient Multi-Task Information Extraction System with Schema-Driven Interfacecited by 9
  2. 2026The Million-Label NER: Breaking Scale Barriers with GLiNER bi-encodercited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

Similar, not cited

Close in wording, no citation either way

  1. 2026GLiGuard: Schema-Conditioned Classification for LLM Safeguard
  2. 2026ProbGuard: Calibrated Safety Risk Estimation from LLM Output Distributions
  3. 2025Do You Really Need a GPU to Guard Your LLM? CPU-Class Classifiers and Multi-Stage Pipelines for Safety Enforcement at Scale
  4. 2026BELLS-O: Evaluating the Operational Trade-offs of LLM Supervision Systems
  5. 2026GRP-Obliteration: Unaligning LLMs With a Single Unlabeled Prompt
  6. 2025Guard Vector: Beyond English LLM Guardrails with Task-Vector Composition and Streaming-Aware Prefix SFT
  7. 2026SingProbe Technical Report
  8. 2026Predict, Don't React: Value-Based Safety Forecasting for LLM Streaming