ParanoiaEval:评测智能编码代理不必要的防御性工作
Get the story
2026年10月7日,研究者在arXiv发布ParanoiaEval,这是首个统一评测编码智能体风险处置能力的基准。该基准基于Avoidance-Transfer-Mitigation-Acceptance框架,包含200个证据控制的仓库级任务对。实验覆盖8个代表性模型,结果显示:尽管有明确证据,仍有11.2%至58.7%的运行出现不必要的风险处置;同时发现更强的任务能力并不保证更恰当的风险处置。
Generated from reports · updated 3 hr ago
Timeline
Follow the coverage from different angles.
- arXiv · Software EngineeringParanoiaEval:评测智能体编码中不必要的防御性工作
研究者推出 ParanoiaEval,首个统一评测编码智能体风险处置能力的基准,基于 Avoidance-Transfer-Mitigation-Acceptance 框架,包含 200 个证据控制的仓库级任务对。实验覆盖 8 个代表性模型,发现尽管有明确证据,11.2%-58.7% 的运行仍出现不必要的风险处置,且更强的任务能力并不保证更恰当的风险处置。
Heat trend
Current heat 9·Comparable peak 10(Oct 7)·Comparable change over 24 hours –
The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.