Add SWE-Bench Verified evaluation results

This PR ensures your model shows up at [https://huggingface.co/datasets/SWE-bench/SWE-bench_Verified.

This is based on the new evaluation results feature: https://huggingface.co/docs/hub/eval-results.

.eval_results/swe_bench_verified.yaml

@@ -0,0 +1,8 @@
+- dataset:
+    id: SWE-bench/SWE-bench_Verified
+    task_id: swe_bench_%_resolved
+  value: 73.1
+  source:
+    url: https://huggingface.co/deepseek-ai/DeepSeek-V3.2/blob/main/assets/paper.pdf
+    name: DeepSeek-V3.2 technical report
+    user: nielsr