English(EN) Can AI draw San Francisco from memory?

AI模型凭记忆绘制旧金山地图，准确率达86%

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-06 05:07

研究人员测试了Anthropic、OpenAI和Google的六款AI聊天模型，要求它们根据经纬度坐标凭记忆识别旧金山的位置。模型在区分水域、公园、建筑物和街道方面的能力得到了评估，表现最好的模型在该城市地图上的准确率达到了86%。尽管街道对大多数模型来说是一个薄弱环节，但有些模型表现出了对密集网格区域与公园区域的理解。成本差异显著，一些表现最佳的模型在相同准确率下比其他模型便宜五倍。 AI

影响展示了LLM的空间推理能力和反向地理编码的潜力，但也突出了街道级别细节的局限性。

排序理由研究论文，详细介绍了测试LLM能力的实验。[lever_c_demoted from research: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 30 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 研究论文，详细介绍了测试LLM能力的实验。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: model release, product; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

dev.to — LLM tag TIER_1 English(EN) · VOYGR · 2026-11-06 05:07

AI能凭记忆画出旧金山吗？

Originally published on <a href="https://voygr.tech/blog/can-ai-draw-san-francisco-from-memory/" rel="noopener noreferrer">voygr.tech</a>. In 2023, <a href="https://arxiv.org/abs/2310.02207" rel="noopener noreferrer">Gurnee and Tegmark</a> showed that you can r…

本文如何被排名

报道来源 [1]

AI能凭记忆画出旧金山吗？

相关实体

相关话题