[KongchangAI]
BenchmarkAsk-F1指标

Ask-F1

A comprehensive metric designed to evaluate proactive questioning strategies; it decomposes "when to ask" and "what to ask" into matching problems along the temporal and content dimensions and combines them via a harmonic mean.

Timeline (last 90 days)

Oct 30

The paper proposes the Ask-F1 metric, which comprehensively measures model performance on "when to ask" and "what to ask".

Unverified50%
Oct 30

Experimental results show that the strongest model reaches only 52.18% on the Ask-F1 metric.

Unverified50%

All Facts (2)

Source Articles