addressing some of the comments

config/ppo/Sorter_curriculum.yaml

           min_lesson_length: 100
           threshold: 0.05
         value: 2.0
-      - name: Lesson1 # The '-' is important as this is a list
+      - name: Lesson1
         completion_criteria:
           measure: progress
           behavior: Sorter
         value: 4.0
-      - name: Lesson2 # This is the start of the second lesson
+      - name: Lesson2
         completion_criteria:
           measure: progress
           behavior: Sorter
           min_lesson_length: 100
           threshold: 0.2
         value: 8.0
-      - name: Lesson4 # The '-' is important as this is a list
+      - name: Lesson4
         completion_criteria:
           measure: progress
           behavior: Sorter
         value: 10.0
-      - name: Lesson5 # This is the start of the second lesson
+      - name: Lesson5
         completion_criteria:
           measure: progress
           behavior: Sorter
           min_lesson_length: 100
           threshold: 0.35
         value: 14.0
-      - name: Lesson7 # This is the start of the second lesson
+      - name: Lesson7
         completion_criteria:
           measure: progress
           behavior: Sorter

docs/Learning-Environment-Design-Agents.md

 
 ### Variable Length Observations
 
-It is possible for agents to collect observations from a varying number of GameObjects by using a `BufferSensor`.
+It is possible for agents to collect observations from a varying number of
+GameObjects by using a `BufferSensor`.
-attention to a varying number of entities. On the trainer side, the `BufferSensor`
+attention to a varying number of entities (for example, a varying number of
+tiles in the [Sorter environment](Learning-Environment-Examples.md#sorter)).
+On the trainer side, the `BufferSensor`
-a flat vector observation. However, attention mechanisms enable solving problems that require comparative reasoning between entities in a scene
-such as our [Sorter environmentt](Learning-Environment-Examples.md#sorter).
+a flat vector observation. However, attention mechanisms enable solving
+problems that require comparative reasoning between entities in a scene
+such as our [Sorter environment](Learning-Environment-Examples.md#sorter).
-the padded observations.
+the padded observations. Note that attention layers are invariant to
+the order of the entities, so there is no need to properly "order" the
+entities before feeding them into the `BufferSensor`.
 
 The `BufferSensor` constructor and Editor inspector have two arguments:
  - `Observation Size` : This is how many floats each entities will be