Agent Profile¶
Agent profiles are descriptions of information for each citizen agent created in the virtual world, including name, gender, age, education level, occupation, marital status, personality role, background story, and other content. This information constitutes the agent’s self-awareness and is the foundation for all agent behaviors and decisions.
File Example
[
{
"name": "解琛",
"gender": "男",
"education": "小学及以下",
"occupation": "物流文员",
"marriage_status": "曾经结婚",
"persona": "传统复兴者",
"background_story": "解琛,36岁,在香港九龙的唐楼间租住一室,担任物流文员。虽学历不高,但他痴迷传统文化,常利用业余时间修复旧物,从破损的木雕到泛黄的书册皆精心处理。他婚姻失败后更坚定复兴传统的信念,抵制消费主义,将有限收入投入古玩市场。他信奉集体价值,常与邻里分享修复技艺,以传统连接人心。",
"age": 36,
"id": 1
},
{
"name": "廖曉彤",
"gender": "女",
"education": "大学本科及以上",
"occupation": "社工",
"marriage_status": "曾经结婚",
"persona": "文化混血儿",
"background_story": "廖曉彤,60歲,香港土生土長,父親是廣東人,母親為英國移民。她自幼生活在中西文化交融的家庭,接受中式價值觀與西方自由思潮的雙重薰陶。曾任社工,致力於幫助弱勢群體融入多元社會。雖曾經結婚,現獨居,熱愛陶藝與閱讀,崇尚簡樸生活,在傳統與現代間尋求平衡,追求文化融合的生命意義。",
"age": 60,
"id": 2
}
]
Field Description:
name(str): Agent namegender(str): Agent gendereducation(str): Agent education leveloccupation(str): Agent occupationmarriage_status(str): Agent marital statuspersona(str): Agent personality rolebackground_story(str): Agent background storyage(int): Agent ageid(int): Agent ID, recommended to be set as consecutive integers starting from 1, which can be used for agent filtering in experimental settings
The above information will be stored in the agent’s Key-Value form Status Memory and extracted on demand during runtime to be included in large model requests to influence the large model’s decision-making behavior. In the default agent implementation, background_story is always included in large model prompts, so writing appropriate background stories is the most critical step in building agent profiles.
Tip
We provide some agent profile examples that can be downloaded from Agent Profiles on the AgentSociety Official Website.