data: add benchmark result lemer-bf16-math-think-temp1.json

results/lemer-bf16-math-think-temp1.json

@@ -0,0 +1,24 @@
+{
+  "model": "/Users/snider/.cache/huggingface/hub/models--lthn--lemer/snapshots/083106e3de02de3606452d58fa0ee2fcb2faff92",
+  "benchmark": "MMLU-Pro",
+  "method": "rapid-mlx + openai SDK + google parse_response",
+  "temperature": 1.0,
+  "processor": "google/gemma-4-E2B-it",
+  "num_runs": 1,
+  "limit_per_subject": 20,
+  "runs": [
+    {
+      "math": {
+        "correct": 12,
+        "total": 20,
+        "accuracy": 0.6
+      },
+      "_overall": {
+        "correct": 12,
+        "total": 20,
+        "accuracy": 0.6
+      },
+      "_elapsed_s": 209.0
+    }
+  ]
+}