Model card: fix table and message block layout

README.md

@@ -180,10 +180,10 @@ The mouse line, copied verbatim:
 - Mouse X Y -> aim / rotate the camera (X>0 right, Y>0 down)
 ```
 
-> **Note.** Lines such as `**Game Rules**`, `**Output Format**` and
-> `**Explanation**` are plain-text section headers that the model saw in training. They are part
-> of the prompt, not Markdown formatting: send them to the model unchanged, including the
-> asterisks.
+**Note.** Lines such as `**Game Rules**`, `**Output Format**` and
+`**Explanation**` are plain-text section headers that the model saw in training. They are part
+of the prompt, not Markdown formatting: send them to the model unchanged, including the
+asterisks.
 
 ### Complete example: Doom Battle-1
 
@@ -301,7 +301,7 @@ The assistant turn is always prefilled, and the request continues it
 | Mode | Prefill | What the model does |
 |---|---|---|
 | `PolicyConfig()` (default, `reasoning=True`) | `<think>` | Decides whether and how long to think, closes `</think>`, then outputs the action |
-| `PolicyConfig(reasoning=False)` | `<think>\n\n</think>` | Thinking is closed and empty; the model outputs `<|action_start|>...<|action_end|>` directly |
+| `PolicyConfig(reasoning=False)` | `<think>\n\n</think>` | Thinking is closed and empty; the model outputs the action chunk directly |
 
 `<think>\n\n</think>` is exactly how the chat template renders a turn without reasoning, so the
 no-reasoning mode stays in the training format. Do not send the request without a prefill.
@@ -311,11 +311,11 @@ Each user turn is the text `current game screen` followed by the frame. Multi-tu
 
 ```text
 system     system prompt (§2)
-user       "current game screen" + frame t-k          ┐ last history_len turns (default 5)
-assistant  action taken at t-k (no reasoning)          ┘
+user       "current game screen" + frame t-k     ┐ last history_len turns
+assistant  action taken at t-k, no reasoning      ┘ (default 5)
 ...
 user       "current game screen" + current frame
-assistant  "<think>"  or  "<think>\n\n</think>"       (prefill, continued by the model)
+assistant  prefill: "<think>" or "<think>\n\n</think>"
 ```
 
 Past turns keep only the action (`history_think=0`). A turn that failed to parse is stored as an