# M6 Attention Maps We show two attention maps of a head using two fixed utterances. The left column shows the maps from the shorter utterance; the right column shows the longer. We also show the properties of a head: **id**, **category**, **globalness (G)**, **verticality (V)**, **Diagonality (D)**. Each metric value of a head is followed by the metric rank among all heads. For example, `G: 5.120 (8)` means the head has the globalness of **5.120** and is the **8th** global head among all heads. ## Layer 1 ![](https://i.imgur.com/okVtuIu.png) ## Layer 2 ![](https://i.imgur.com/shCJ7uD.png) ## Layer 3 ![](https://i.imgur.com/UjcYVgX.png) ## Layer 4 ![](https://i.imgur.com/GVueK5I.png) ## Layer 5 ![](https://i.imgur.com/VZp3vBA.png) ## Layer 6 ![](https://i.imgur.com/MbvJvF9.png) ## Layer 7 ![](https://i.imgur.com/e6w07FU.png) ## Layer 8 ![](https://i.imgur.com/REXdQr7.png) ## Layer 9 ![](https://i.imgur.com/uxtRjtT.png) ## Layer 10 ![](https://i.imgur.com/9LDf2MQ.png) ## Layer 11 ![](https://i.imgur.com/2byQjNk.png) ## Layer 12 ![](https://i.imgur.com/FPPsBi8.png)