Add EvasionBench evaluation results

## Summary

Adding evaluation results from EvasionBench benchmark for managerial evasion detection in earnings call Q&A.

## Details

- **Benchmark**: [EvasionBench](https://huggingface.co/datasets/FutureMa/EvasionBench)
- **Paper**: [arXiv:2601.09142](https://arxiv.org/abs/2601.09142)
- **Metric**: Macro-F1
- **Score**: 78.16%

## Reference

For more details about the evaluation methodology, please refer to:
- [EvasionBench Dataset](https://huggingface.co/datasets/FutureMa/EvasionBench)
- [Project Page](https://iiiiqiiii.github.io/EvasionBench)

.eval_results/evasionbench.yaml

@@ -0,0 +1,8 @@
+- dataset:
+    id: FutureMa/EvasionBench
+    task_id: evasion_bench
+  value: 78.16
+  date: "2026-02-10"
+  source:
+    url: https://arxiv.org/abs/2601.09142
+    name: EvasionBench Paper