On HotpotQA with Contriever augmentation and a HotpotQA-retrained predictor, targeted intervention improves accuracy from 42.2% to 52.2%, essentially matching the 52.3% oracle result.

On HotpotQA with Contriever augmentation and a HotpotQA-retrained predictor, targeted intervention improves accuracy from 42.2% to 52.2%, essentially matching the 52.3% oracle result. · CiteArk

测量结果

已评估 0/10 项

获得支持 0有争议或冲突 0尚未评估 10

保留同一结论下的完整结果组。已有评估排在前面；单项判断不替代整组结论。展开行查看条件、来源与历次证据。

每行是一项测量，分别显示论文报告、最近可用实测和累计证据状态。
指标与条件	论文报告	最近可用实测	证据判断
condition: Baseline	42.2%	—	尚未评估
condition: Baseline	0%	—	尚未评估
condition: Baseline	0 calls/question	—	尚未评估
condition: DeBERTa targeted	52.2%	—	尚未评估
condition: DeBERTa targeted	53.4%	—	尚未评估
condition: DeBERTa targeted	1.07 calls/question	—	尚未评估
condition: Oracle	52.3%	—	尚未评估
condition: Oracle	54.1%	—	尚未评估
condition: Oracle	1.08 calls/question	—	尚未评估
	10 percentage points	—	尚未评估

On HotpotQA with Contriever augmentation and a HotpotQA-retrained predictor, targeted intervention improves accuracy from 42.2% to 52.2%, essentially matching the 52.3% oracle result.

报告指标与观测值

Assessment（0）

关联运行（0）

测量结果