Tried it with N=2 and M=1 (brother singular) with the gpt-4o model and CoT. 1. 50% success without "full" terminology. 2. 5% success with "full" terminology.

Tried it with N=2 and M=1 (brother singular) with the gpt-4o model and CoT. 1. 50% success without "full" terminology. 2. 5% success with "full" terminology. So, the improvement in clarity has exactly the opposite effect.