Files
myWiki/concepts/evaluator-metrics.md

43 lines
1.4 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

---
title: "评估器质量指标Evaluator Metrics"
created: 2026-07-02
updated: 2026-07-02
type: concept
tags: [evaluation, metrics, rft, rl, benchmarking]
sources:
- "[[verification-horizon-no-silver-bullet]]"
---
# 评估器质量指标
衡量 [[agent-evaluator|Agent 评估器]] 与 ground-truth单元测试分数对齐程度的多维度指标框架。
## 指标体系
| 指标 | 衡量什么 | 使用场景 |
|------|---------|---------|
| **BoN Accuracy** | 评估器能否从候选池中选出最佳样本 | RFT小候选池 |
| **Regret** | 次优选择的质量损失 | 联合评估 |
| **Kendall's τ** | 全分数范围的排名一致性 | RL需要细粒度梯度 |
| **Pearson r / Spearman ρ** | 宏观平均相关性 | 综合评估 |
| **Threshold-Conditioned UT Score** | 过滤后样本的平均质量 | RFT大候选池 |
## 指标冲突
**排名能力 ≠ 过滤质量**
- Qwen 3.7 Plus BoN-Acc (67.3%) > DeepSeek V4 Pro (54.5%)
- 但 DeepSeek 的 θ≥8 条件下 UT score (0.611) > Qwen (0.595)
## 训练目标与指标映射
| 训练目标 | 关键指标 | 关注点 |
|---------|---------|--------|
| RFT候选充足 | Threshold-Conditioned UT Score | 低 false positive |
| RFT候选有限 | UT Score + retained count | 平衡 FP/FN |
| RL | Kendall τ + score discrimination | 细粒度梯度 |
## 参考
- [[verification-horizon-no-silver-bullet|论文原文]]
- [[agent-evaluator|Agent 评估器]]