Skip to content
Hot eventLive

仅结论后训练小模型证据恢复研究

1 reports1 sources3 hr ago updated

Get the story

AI overview

该研究在 ContractNLI 上衡量小语言模型仅凭结论后训练时能恢复多少证据,人类证据片段直到评估才放出。研究指出,匹配结论与吻合证据片段并不等同:六个系统中两项分数仅弱相关且排序不同,accuracy 不足以指导需可审查引用的场景。

Generated from reports · updated 3 hr ago

Timeline

Follow the coverage from different angles.

Oct 7, 2026
  1. arXiv · Computation and Language
    仅有结论而无标注证据:拒绝采样还是仅标签后训练用于证据恢复?

    研究在 ContractNLI 上衡量小语言模型仅凭结论后训练时能恢复多少证据,人类证据片段直到评估才放出。匹配结论与吻合证据片段并不等同:六个系统中两项分数仅弱相关且排序不同,accuracy 不足以指导需可审查引用的场景。

Heat trend

Current heat 9·Comparable peak 10(Oct 7)·Comparable change over 24 hours –

02.557.510Oct7Oct7Oct7Oct7

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.