I'm not surprised that llama 3.2 11b failed here. The author should really have tested Llama 3.2 90b, which does remarkably better:
Same query and image:
This line graph illustrates the working-poor rate of the US labour force from 1986 to 2022 (those working and below the designated poverty line). The graph can be broken down into 3 main proceedings.
1. The first spike in working-poor individuals was displayed in 1992.
2. That peak was then over shadowed by a dramatic incline and following higher peak in 2010 possibly caused by the 2008 housing crisis
3. 2010 was then followed almost immediately by a steady decline to the point of bringing the working-poor rate below 4 percent for the first time in 30 years in 2022.
I'm not surprised that llama 3.2 11b failed here. The author should really have tested Llama 3.2 90b, which does remarkably better:
Same query and image:
This line graph illustrates the working-poor rate of the US labour force from 1986 to 2022 (those working and below the designated poverty line). The graph can be broken down into 3 main proceedings. 1. The first spike in working-poor individuals was displayed in 1992. 2. That peak was then over shadowed by a dramatic incline and following higher peak in 2010 possibly caused by the 2008 housing crisis 3. 2010 was then followed almost immediately by a steady decline to the point of bringing the working-poor rate below 4 percent for the first time in 30 years in 2022.