Large language models (LLMs) are becoming foundational to email security products. Increasingly, they power classification, threat detection, triage, and analyst workflows. While hundreds of public benchmarks measure general reasoning and cybersecurity, none are designed to measure whether an LLM understands email communication and email security as first-class domains.
emailbench fills that gap. It is a compact, high-quality benchmark for email understanding, built… See the full description on the dataset page:
https://huggingface.co/datasets/proofpoint-ai/emailbench.
