原帖引用 · AI 翻译
标题：“谁应该知道什么？面向多智能体大语言模型经济体的实证信息设计”
斯坦福大学的一篇新论文表明，更多信息会帮助一些大语言模型智能体更好地交易，却会损害另一些智能体的交易表现，具体取决于模型。
当参与拍卖的买卖双方 AI 智能体被明确告知各方对商品的估值时，Claude Opus 5.5 获得的潜在收益份额从 94% 降至 64%。
多智能体系统需要规则来规定每个智能体能看到有关其他智能体的哪些信息，例如确切价格，或仅有排名。通常的假设是，更多信息会有所帮助。
他们让 10 个前沿大语言模型参与交易、议价和双向拍卖，仅改变私有估值的共享精度，并根据已知效用评估所得收益。
在交易中，披露确切估值使 GPT-6 Sol 的收益从最高可能收益的 98% 降至 91%，却使 Claude Sonnet 5 的收益从 72% 升至 86%。在拍卖中，完全披露推高了要价，并使 Grok 4.7 的收益从 97% 降至 62%。
当智能体代表你谈判时，应将模型及其获得的信息结合起来测试，因为更多细节会让一些模型变得更贪婪。
原文
New Stanford paper shows that more information helps some LLM agents trade better and hurts others, depending on the model. AI agents buying and selling in an auction were told exactly how much each side valued the goods, Claude Opus 5.5’s share of possible gains fell from 94% to 64%. Multi-agent systems need rules on what each agent sees about others, such as exact prices or only rankings. The usual assumption is that more information helps. They ran 10 frontier LLMs in trading, bargaining, and double auctions, changing only how precisely private values were shared, and scored gains against known utilities. In trade, exact values cut GPT-6 Sol's gains from 98% to 91% of the maximum but raised Claude Sonnet 5's from 72% to 86%. In auctions, full disclosure inflated price demands and dropped Grok 4.7 from 97% to 62%. When agents negotiate for you, test the model and what it gets told together, because more detail makes some models greedier.