model card: Thai showcase poster first, Doom demo after

README.md

@@ -80,34 +80,16 @@ from transformers import AutoModel, AutoTokenizer
 model = AutoModel.from_pretrained("iapp/OpenThai-SystemOne", trust_remote_code=True)
 ```
 
-## Demo: playing Doom, no vision, no text
+## Demos
 
-![OpenThai-SystemOne playing Doom](https://huggingface.co/iapp/OpenThai-SystemOne/resolve/main/assets/doom_10s.gif)
+### Thai showcase (real v0.3 outputs via the API)
 
-The model controls a Doom marine ([ViZDoom](https://github.com/Farama-Foundation/ViZDoom)) in real time. Every 4 game
-tics the engine's symbolic state is serialised to text, for example:
+![OpenThai-SystemOne Thai showcase: 7 everyday Thai tasks with the model's real probabilities](https://huggingface.co/iapp/OpenThai-SystemOne/resolve/main/assets/thai-showcase.png)
 
-```text
-health 100/100 | ammo 50 | kills 1
-crosshair: empty; nearest visible enemy 39deg to the RIGHT
-enemies: Zombieman 5m right -39deg VISIBLE; ChaingunGuy 19m right -24deg; Zombieman 19m ahead -12deg
-items: GreenArmor 41m right -18deg
-depth ahead: 35/255 (obstacle near)
-last actions: ATTACK ATTACK ATTACK ATTACK
-```
+Seven everyday Thai tasks, each answered in one forward pass; probabilities are the model's actual output. Example 1 keeps a real miss (politeness) on purpose.
 
-and the model answers two typed questions in one forward pass: a `choice` over the 7 actions (the key that gets
-pressed) and a `noul` "is an enemy in the crosshair". About 41 ms per decision on one H100 (~24 decisions/s),
-0 output tokens, and no Doom data in training: everything comes from reading the state and the option descriptions.
-It misses shots and dies on hard levels; the point is the speed and the calibrated probabilities, the same mechanics
-that route tickets or pick UI elements for an agent.
-
-Run it on your machine (iApp API key or the local weights, live HUD in Thai or English, optional recording):
-**https://github.com/iapp-technology/openthai-systemone-doom**
-
-## Thai showcase (real v0.3 outputs, 2026-09-22, via the API)
-
-Seven everyday Thai tasks, each answered in one forward pass; probabilities are the model's actual output.
+<details>
+<summary>Full inputs, questions and outputs of the 7 examples</summary>
 
 **1. Support ticket triage** — state: *"แอปโอนเงินไม่ได้ตั้งแต่เมื่อคืน ขึ้นว่า error 502 ตลอด ลองลงใหม่แล้วก็ยังไม่หาย รบกวนช่วยด่วนนะครับ ต้องโอนค่าเทอมลูกพรุ่งนี้"*
 
@@ -140,6 +122,33 @@ Note on example 1's politeness flag: the ticket uses "รบกวน…นะ
 at 99%. Politeness judgement on formal Thai is a known gap and will get a targeted set in v0.4. We keep the miss here
 because a showcase that only shows hits is not useful.
 
+</details>
+
+### Playing Doom, no vision, no text
+
+![OpenThai-SystemOne playing Doom](https://huggingface.co/iapp/OpenThai-SystemOne/resolve/main/assets/doom_10s.gif)
+
+The model controls a Doom marine ([ViZDoom](https://github.com/Farama-Foundation/ViZDoom)) in real time. Every 4 game
+tics the engine's symbolic state is serialised to text, for example:
+
+```text
+health 100/100 | ammo 50 | kills 1
+crosshair: empty; nearest visible enemy 39deg to the RIGHT
+enemies: Zombieman 5m right -39deg VISIBLE; ChaingunGuy 19m right -24deg; Zombieman 19m ahead -12deg
+items: GreenArmor 41m right -18deg
+depth ahead: 35/255 (obstacle near)
+last actions: ATTACK ATTACK ATTACK ATTACK
+```
+
+and the model answers two typed questions in one forward pass: a `choice` over the 7 actions (the key that gets
+pressed) and a `noul` "is an enemy in the crosshair". About 41 ms per decision on one H100 (~24 decisions/s),
+0 output tokens, and no Doom data in training: everything comes from reading the state and the option descriptions.
+It misses shots and dies on hard levels; the point is the speed and the calibrated probabilities, the same mechanics
+that route tickets or pick UI elements for an agent.
+
+Run it on your machine (iApp API key or the local weights, live HUD in Thai or English, optional recording):
+**https://github.com/iapp-technology/openthai-systemone-doom**
+
 ## Evaluation
 
 All numbers are zero-shot: the model sees only the state, the instructions and the option names/descriptions.