按年被提到
20262026 · 2 次2026
AI 安全、伦理与社会影响›大模型与人类行为·arXiv:2603.23841 · cs.CL

PoliticsBench: Benchmarking Political Values in Large Language Models with Multi-Turn Roleplay

Rohan Khetan, Ashna Khetan · 2026
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2026-04-01 · 最近 2026-07-08
2不同的人
2活跃月份
2讨论串
0被引(OpenAlex)
arXiv 摘要(候选层不写导读)

While Large Language Models (LLMs) are increasingly used as primary sources of information, their potential for political bias may impact their objectivity. Existing benchmarks of LLM social bias primarily evaluate demographic stereotypes, and when political bias is measured, it is done so at a coarse level, overlooking the values that shape sociopolitical reasoning. We introduce PoliticsBench, a multi-stage roleplay benchmark for evaluating fine-grained value expression in LLMs. Across twenty evolving scenarios, models articulate tradeoffs, take positions, and make decisions under competing pressures. Across eight prominent LLMs, we show that scenario-based prompting elicits broader and more strongly expressed value profiles than direct political questions, with peak interaction stages increasing the number of strongly activated value dimensions by approximately $0.75$ (out of 10 total dimensions), a statistically significant increase relative to baseline prompting ($p < 0.05$). In addition, commitment to a stance increases over the course of interaction, rising by approximately $1.4$ points on a $[0,5]$ scale from initial to decision stages. While responses become less robust to scenario paraphrasing in later interaction stages, inter-judge agreement remains relatively stable. Our results suggest that evaluating LLM political behavior requires moving beyond static prompts toward longer interactive settings that capture how values are applied in context.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论WatchDog · 2026-04

“Overall it's worse than the other frontier models, but it's decent for queries about breaking news, due to being trained on twitter data. It's also better for queries about controversial topics, and topics that the other labs have deemed to be "unsafe". Politically, it differs quite a bit from other models.[0] It's right leaning, although it's closer ne…”

SpaceX Files to Go Public原评论 ↗
评论FunHearing3443 · 2026-07

“Studies on political bias in models consistently show that LLMs lean politically left. The only outlier is grok which leans right but by a smaller factor, according to this study for instance: https://arxiv.org/abs/2603.23841”