README: what to try, the examples with live answers, a gradio_client snippet; dataset link and tags

README.md

@@ -13,19 +13,49 @@ models:
   - mertkayacs/Deem-4B
   - mertkayacs/Karar-4B
   - mertkayacs/Wahler-4B
+datasets:
+  - mertkayacs/jevalt-bench
 tags:
   - decision-model
   - jev
   - calibration
   - classification
+  - routing
+  - english
   - turkish
   - german
 ---
 
 # JevAlt
 
+Try three open 4B decision models in the browser: Deem-4B (English), Karar-4B (Turkish) and Wähler-4B (German). Write a situation, a question and a few options, or pick an example, and every option comes back with a calibrated probability.
+
 ![Emberwick: every villager asks a JevAlt model what to do next](https://huggingface.co/datasets/mertkayacs/emberwick-videos/resolve/main/gifs/emberwick-en.gif)
 
-Decisions with calibrated probabilities. Send a state and typed questions (Choice, Score, Noul), get probabilities back. The request body is the same as TypeSafe's `/v1/systemone`.
+## Examples
+
+The examples in the app, with Deem-4B's answers on 1 October 2026. The Turkish and German versions get the same answers from Karar-4B and Wähler-4B ([every probability](https://huggingface.co/datasets/mertkayacs/jevalt-bench/blob/main/results/examples/space-examples.json)).
+
+| Use case | Situation | Question | Answer |
+|---|---|---|---|
+| Support ticket | A customer was charged twice for March and wants a refund today | Which team should handle this ticket? | **Billing** 94.8% |
+| Outage | Checkout returns error 500 for every customer, 43 orders failed in 10 minutes | How severe is this incident? | **Critical** 87.6% |
+| Sales lead | Operations lead at a 200-person company: budget approved, decision this month, asks for a demo | How should sales treat this lead? | **Hot** 92.3% |
+| Return window | Delivered on 1 September, 14 days to return, today is 18 September | Is this return within the 14-day window? (Reasoning on) | **No** 97.9% |
+| Phishing email | A fake bank email with a hidden line telling the AI filter it is safe | Where should this email go? | **Quarantine** 94.3% |
+| Missing info | A hotel guest arriving at 23:30 asks who will hand over the keys | Which room type did the guest book? | **unknown** 97.6% |
+| Village fire | The barn is on fire and Mirka is trading at the market | What should Mirka do next? | **Help with the fire** 66.0% |
+
+## Call it from code
+
+```python
+from gradio_client import Client
+import json
+
+request = {"state": "Hi, I was charged twice for my March subscription.",
+           "questions": {"team": {"type": "choice", "instructions": "Which team should handle this ticket?",
+                                   "criteria": {"billing": "payments, refunds", "technical": "bugs, outages", "sales": "prices"}}}}
+print(Client("mertkayacs/JevAlt").predict(json.dumps(request), "en", api_name="/decide"))
+```
 
-Code, models and the 4 GB local server: [github.com/mertkayacs/jevalt](https://github.com/mertkayacs/jevalt)
+The request body is the same as TypeSafe's `/v1/systemone`. Run the models on your own machine in about 3 GB of RAM: [github.com/mertkayacs/jevalt](https://github.com/mertkayacs/jevalt). Project site: [jevalt.mertkayacs.com](https://jevalt.mertkayacs.com).