第三章 · Section 3
General Knowledge (K) · 常识
General Knowledge (K): Knowledge that is familiar to most well-educated people or is important enough that most adults have been exposed to it.
常识（K）：大多数受过良好教育的人熟悉、或重要到大多数成年人接触过的知识。
Background knowledge about how the world works.
Sample tasks: "What happens if you drop a glass bottle on concrete?" / "Does making a sandwich take longer than baking bread?"
关于世界如何运转的背景知识。
样例问题："把玻璃瓶掉在混凝土上会发生什么？" / "做三明治比烤面包花的时间更长吗？"
Knowledge of natural and physical sciences.
Physics sample: "A 2 kg object moves at constant velocity of 3 m/s. What is the net force?"
Chemistry sample: "State the molecular geometry for the sulfur tetrafluoride molecule."
Biology sample: "Which molecule is the final electron acceptor in the electron transport chain of cellular respiration?"
自然与物理科学的知识。
物理样例："一个 2 kg 的物体以 3 m/s 的恒定速度运动。净力是多少？"
化学样例："说出四氟化硫分子的分子几何构型。"
生物样例："细胞呼吸电子传递链中，最终的电子受体是哪个分子？"
Understanding of human behavior, societies, and institutions.
Psychology sample: "Name the Big Five personality traits."
Microeconomics sample: "Define a positive externality."
Macroeconomics sample: "What's the difference between nominal and real interest rates?"
Geography sample: "Which continent is the most threatened by desertification?"
Comparative Government sample: "Describe the role of the Guardian Council in Iran."
对人类行为、社会与制度的理解。
心理学样例："说出大五人格特质。"
微观经济学样例："定义正外部性。"
宏观经济学样例："名义利率与实际利率有什么区别？"
地理学样例："哪个大洲受荒漠化威胁最大？"
比较政治学样例："描述伊朗监护委员会的角色。"
Knowledge of past events and objects.
European History sample: "What were the main goals of the Congress of Vienna in 1815?"
US History sample: "Analyze the goals of the Civil Rights Movement of the 1950s."
World History sample: "Describe the end of the Cold War."
Art History sample: "Discuss the use of contrapposto in ancient Greek and Roman sculpture."
对过去事件与事物的知识。
欧洲史样例："1815 年维也纳会议的主要目标是什么？"
美国史样例："分析 1950 年代民权运动的目标。"
世界史样例："描述冷战结束。"
艺术史样例："讨论对跖站姿（contrapposto）在古希腊罗马雕塑中的运用。"
Cultural literacy and awareness.
Current Affairs sample: "Who's the president of the United States?"
Popular Culture sample: "Who is this?"
文化素养与文化意识。
时事样例："美国总统是谁？"
流行文化样例："这是谁？"
评估细节与 AI 表现
Assessment Details & AI Performance
Assessment Details. See Appendix A for further details on how to assess general knowledge capabilities concretely.
AI System Performance. The table summarizes current AI system performance on General Knowledge (K) tasks. GPT-4 has substantial general knowledge, and GPT-5 partially fills in its remaining gaps.
Model | Commonsense (2%) | Science (2%) | Social Science (2%) | History (2%) | Culture (2%) | Total
GPT-4 | 2% | 2% | 2% | 2% | 0% | 8%
GPT-5 | 2% | 2% | 2% | 2% | 1% | 9%
评估细节：如何在具体层面评估常识能力，参见附录 A。
AI 系统表现：下表汇总当前 AI 系统在常识（K）任务上的表现。GPT-4 已有相当扎实的常识，GPT-5 部分填补了它剩余的缺口。
模型 | 常识性知识(2%) | 科学(2%) | 社会科学(2%) | 历史(2%) | 文化(2%) | 合计
GPT-4 | 2% | 2% | 2% | 2% | 0% | 8%
GPT-5 | 2% | 2% | 2% | 2% | 1% | 9%