Add community evaluation results for HLE

This PR adds community-provided evaluation results for the following benchmarks:

- **[HLE](https://huggingface.co/datasets/cais/hle)**

These results were extracted from the model card. This is based on the new [evaluation results feature](https://huggingface.co/docs/hub/eval-results).*Note: This is an automated PR. Please review the evaluation results before merging.*

.eval_results/hle.yaml

@@ -0,0 +1,6 @@
+- dataset:
+    id: cais/hle
+  value: 19.1
+  source:
+    url: https://huggingface.co/openbmb/AgentCPM-Explore
+    name: Model Card