JevBench adapter: report input token usage

README.md

@@ -63,4 +63,5 @@ asks from the command line.
 
 `examples/jevbench_adapter.py` is an adapter for the
 [JevBench](https://github.com/fstandhartinger/jevbench) harness: it sends the harness's own question
-to `system_one` and returns the probabilities.
+to `system_one` and returns the probabilities, with the input tokens the model read as `usage`
+(output tokens are always 0).

examples/jevbench_adapter.py

@@ -43,6 +43,7 @@ class YevAdapter:
         else:
             answer = response.choices["decision"] if kind == "choice" else response.scores["decision"]
             result.probs = {str(k): float(v) for k, v in answer.probabilities.items()}
+        result.usage = {"input_tokens": response.usage.input_tokens, "output_tokens": 0}
         result.ok = True
         result.raw = {"runtime": {"probability_origin": "native-softmax", "windows": "mean"}}
         return result