Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

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

About

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.

Bogdan Minko, Sabrina Sadiekh, Evgeniy Kokuykin• 2026

Related benchmarks

TaskDatasetResultRank
Safety ModerationStrongREJECT
F1 Score99.7
15
Safety ModerationPolyGuard
Prompt F171.7
15
Safety ModerationAegis 2.0
Prompt F180.2
15
Safety ModerationAegis, StrongReject, PolyGuard Aggregate 2.0
Average F1 Score76.9
15
Safety ClassificationAegisSafetyTest V2--
14
Binary Safety ClassificationToxicChat jailbreaking
Macro F170.54
11
Binary Safety Classificationoai_safety OpenAI moderation
Macro F167.85
11
Inference Efficiency1024-token sequences (inference summary)
Throughput (Samples/s)34.49
11
Binary Safety Classificationwildguard prompt safety
Macro F172.62
11
PII detectionPII-Bench
Name F183.1
10
Showing 10 of 28 rows

Other info

Follow for update