Skip to content
Hot eventLive

LLM 否定处理机制与基准研究

1 reports1 sources4 hr ago updated

Get the story

AI overview

2026-10-08，arXiv Computation and Language 发布一手研究，在自建否定评测基准上测试近期开源与闭源 LLM，考察模型如何改变否定条件下的预测。结果显示，在 37%–71% 的情况下，模型在否定问句下仍重复原答案（例如“What is not the capital of Spain?”）。该研究为目前唯一报道，事件尚无更早或后续报道，亦无相互矛盾的数字或说法。

Generated from reports · updated 3 hr ago

LatestOct 8
2026-10-08 arXiv 研究称，37%–71% 情况下 LLM 在否定问句下仍重复原答案。

Timeline

Follow the coverage from different angles.

Oct 8, 2026
  1. arXiv · Computation and Language
    LLM 如何改变否定条件下的预测？

    研究在自建的否定评测基准上测试近期开源与闭源 LLM，发现 37-71% 的情况下模型在否定问句下仍重复原答案（如"What is not the capital of Spain?

Heat trend

Current heat 9·Comparable peak 10（Oct 8）·Comparable change over 24 hours –

02.557.510Oct8Oct8Oct8Oct8

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.