← 研踪 PaperTrailarXiv:2611.03939 · v1
From Benchmarks to Behaviors: A Framework for Auditing AI Agents
首次提交 2026-11-03 04:41
版本更新 2026-11-03 04:41
雷达观测 2026-11-04 23:40
摘要
Generative AI systems increasingly mediate access to information. Their behavior depends on tools, instructions, and user memories that are often difficult for researchers to observe or control. Here, I introduce TRACE (Transparent Research on Agent Contexts and Environments), a framework for systematically auditing how these components shape AI agent behavior. Built on an open-source agent harness, TRACE enables controlled manipulation of system components. It also records inputs, outputs, and task execution for downstream analysis. I demonstrate this framework through an experiment examining how user memories influence source selection when agents answer questions about the 2026 U.S. midterm elections. I find that memories expressing explicit preference for a publisher increase its inclusion among the first three cited sources in a response by 33-39 percentage points. Memories identifying the user as a Republican or expressing distrust of mainstream media shift citations rightward. That same media distrust memory also increases citation of sources expressing misinformation or conspiracy theories by 6.3 percentage points, with substantial variation across models. These findings demonstrate that concise user memories can alter source selection in consequential settings, even when questions and search results remain fixed. TRACE extends algorithm auditing to agent harnesses, providing a transparent way to study agent behaviors that benchmarks or commercial interfaces may struggle to isolate.
作者备注：39 pages, 4 figures
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-03 04:41 · 最新观测版本
From Benchmarks to Behaviors: A Framework for Auditing AI Agents
Generative AI systems increasingly mediate access to information. Their behavior depends on tools, instructions, and user memories that are often difficult for researchers to observe or control. Here, I introduce TRACE (Transparent Research on Agent Contexts and Environments), a framework for systematically auditing how these components shape AI agent behavior. Built on an open-source agent harness, TRACE enables controlled manipulation of system components. It also records inputs, outputs, and task execution for downstream analysis. I demonstrate this framework through an experiment examining how user memories influence source selection when agents answer questions about the 2026 U.S. midterm elections. I find that memories expressing explicit preference for a publisher increase its inclusion among the first three cited sources in a response by 33-39 percentage points. Memories identifying the user as a Republican or expressing distrust of mainstream media shift citations rightward. That same media distrust memory also increases citation of sources expressing misinformation or conspiracy theories by 6.3 percentage points, with substantial variation across models. These findings demonstrate that concise user memories can alter source selection in consequential settings, even when questions and search results remain fixed. TRACE extends algorithm auditing to agent harnesses, providing a transparent way to study agent behaviors that benchmarks or commercial interfaces may struggle to isolate.
查看此版本解读 →查看此版本原文 ↗