NVIDIA 论文测试长程智能体可靠性:128K token 任务准确率比 4K 低 62.8%
中文摘要
NVIDIA提出Long-Transduction方法测试长程智能体,发现128K token任务准确率比4K低62.8%。
English Summary
NVIDIA's paper introduces Long-Transduction to test long-horizon agent reliability, finding 128K token task accuracy is 62.8% lower than 4K tasks.
原文节选
NVIDIA 发布论文《Staying on Task: Testing the Foundations of Long-Horizon Agent Reliability》(arxiv.org/abs/2609.38712),提出 Long-Transduction 诊断方法,测试模型在长生成中持续保持任务的能力。