Add development-suite evaluation results for the 8-bit file

README.md

@@ -10,6 +10,34 @@ tags:
 - typed-decisions
 - 8-bit
 - apple-silicon
+model-index:
+  - name: kev-4b-mlx-8bit
+    results:
+      - task: { type: text-classification, name: typed decision, 8-bit MLX }
+        dataset: { type: mixed, name: "transfer-v4 development (656 questions; 8-bit, same Mac as bf16)" }
+        metrics:
+          - { type: accuracy, value: 0.817 }
+          - { type: brier_score, value: 0.243 }
+      - task: { type: text-classification, name: typed decision, 8-bit MLX }
+        dataset: { type: mixed, name: "decision-v7 development (1,264 questions; 8-bit, same Mac as bf16)" }
+        metrics:
+          - { type: accuracy, value: 0.873 }
+          - { type: expected_calibration_error, value: 0.010, name: "ECE, as served" }
+      - task: { type: text-classification, name: typed decision, 8-bit MLX }
+        dataset: { type: mixed, name: "hard-v1 development (1,083 questions; 8-bit, same Mac as bf16)" }
+        metrics:
+          - { type: accuracy, value: 0.786 }
+          - { type: brier_score, value: 0.313 }
+      - task: { type: text-classification, name: typed decision, 8-bit MLX }
+        dataset: { type: mixed, name: "devtools-v1 development (1,074 questions; 8-bit, same Mac as bf16)" }
+        metrics:
+          - { type: accuracy, value: 0.746 }
+          - { type: brier_score, value: 0.357 }
+      - task: { type: text-classification, name: typed decision, 8-bit MLX }
+        dataset: { type: mixed, name: "documents-v1 development (920 questions; 8-bit, same Mac as bf16)" }
+        metrics:
+          - { type: accuracy, value: 0.896 }
+          - { type: brier_score, value: 0.173 }
 ---
 
 # Kev-4B, merged and quantized to 8 bits for MLX