{"slug": "gpf-livenews-a-streaming-evaluation-protocol-for-group-conditioned-framing-in", "title": "GPF-LiveNews: A Streaming Evaluation Protocol for Group-Conditioned Framing in Large Language Models", "summary": "Researchers introduced GPF-LiveNews, a streaming evaluation protocol that audits how large language models frame breaking news for different audience groups. The protocol expands BBC and Reuters news anchors across 42 identity labels and seven prompt families, then measures semantic sensitivity and sentiment disparity in model outputs. A pilot study of 23 models over 12 monitoring runs found that Policy/Action prompts produced the strongest semantic movement, though sentiment variation remained relatively flat across dimensions.", "body_md": "arXiv:2605.28848v1 Announce Type: new\nAbstract: Deployed language models are evaluated in a non-stationary environment: model versions, retrieval layers, safety systems, and real-world inputs all change over time. Static bias benchmarks remain useful, but they do not show how models frame newly emerging events for different prompted audiences. We introduce GPF-LIVENEWS, a streaming evaluation protocol and benchmark snapshot for auditing group-conditioned framing in open-ended LLM outputs. The protocol expands fresh BBC/Reuters news anchors across 42 identity labels and seven prompt families, then evaluates response bundles using semantic-sensitivity and sentiment-disparity signals. In a pilot over 12 monitoring runs and 23 hosted models, Policy/Action prompts produce the strongest semantic movement, while sentiment variation is flatter across dimensions and prompt families. The released artifact includes article metadata, prompt templates, instantiated prompts, model-output metadata, score tables, documentation, and reproduction scripts. We interpret all scores as observed-window audit signals for human review, not as permanent fairness rankings or direct proof of harmful bias.", "url": "https://wpnews.pro/news/gpf-livenews-a-streaming-evaluation-protocol-for-group-conditioned-framing-in", "canonical_source": "https://arxiv.org/abs/2605.28848", "published_at": "2026-05-29 04:00:00+00:00", "updated_at": "2026-05-29 04:25:40.156766+00:00", "lang": "en", "topics": ["large-language-models", "ai-safety", "ai-ethics", "ai-research", "natural-language-processing"], "entities": ["GPF-LiveNews", "BBC", "Reuters"], "alternates": {"html": "https://wpnews.pro/news/gpf-livenews-a-streaming-evaluation-protocol-for-group-conditioned-framing-in", "markdown": "https://wpnews.pro/news/gpf-livenews-a-streaming-evaluation-protocol-for-group-conditioned-framing-in.md", "text": "https://wpnews.pro/news/gpf-livenews-a-streaming-evaluation-protocol-for-group-conditioned-framing-in.txt", "jsonld": "https://wpnews.pro/news/gpf-livenews-a-streaming-evaluation-protocol-for-group-conditioned-framing-in.jsonld"}}