兩項目評估報告
- EvalRun: #209（suite #122
auto-8c188cbc-r1-053959534）
- Target: #1 production-baseline
- Cases: 30（N=1, attempts=30/30）
- Generated: 2026-05-16T13:51:00+08:00
本報告含兩個獨立評估項目。項目一 評估 bot 進入「知識與產品查詢」fallback
後的撈取 + 回答能力；項目二 評估 bot 跨全 enabled scenarios 的 routing →
tool → answer 三段 funnel 健康度。
項目一: 知識庫精準度
評估範圍 (prerequisite: 進對 scenario)
- 進對 scenario 的 attempts: 10 / 30 (33.3%)
- 只對這些 attempts 算 retrieval + answer
- 沒進對 scenario 的 attempts → 排除（routing 失敗不該污染 KB 訊號）
Per-scenario 評估
| Scenario |
qualifying / total attempts |
retrieval_relevance |
answer_correctness |
| 知識與產品查詢 |
10 / 10 |
❌ 38.3% |
✅ 100.0% |
待修方向（worst-3）
- 知識與產品查詢 × retrieval_relevance = 38.3% — 撈到的 KB chunks 不夠相關 — KB content gap，跟 Neptune team 提 enrichment
- 知識與產品查詢 × answer_correctness = 100.0% — 回答內容有問題 — 細看 hallucination_rate vs answer_quality 哪邊低
項目二: 情境調用與完成
整體 funnel（全 scenarios 加總）
| Stage |
Pass count |
% of total |
| Total attempts |
20 |
100.0% |
| Step 1: scenario_routing |
7 |
35.0% |
| Step 2: + tool_calling |
7 |
35.0% |
| Step 3: + answer |
3 |
15.0% |
Per-scenario funnel
| Scenario (n_attempts) |
Step 1 (routing) |
Step 2 (tool) |
Step 3 (answer) |
end-to-end |
| early_return (9) |
❌ 0/9 (0.0%) |
❌ 0/9 (0.0%) |
❌ 0/9 (0.0%) |
❌ 0/9 (0.0%) |
| 用戶反映付款/儲值後未收到鑽石 (1) |
✅ 1/1 (100.0%) |
✅ 1/1 (100.0%) |
✅ 1/1 (100.0%) |
✅ 1/1 (100.0%) |
| 用戶問題回報情境 (3) |
❌ 1/3 (33.3%) |
❌ 1/3 (33.3%) |
❌ 1/3 (33.3%) |
❌ 1/3 (33.3%) |
| 用戶申請開通超商代碼付款、轉帳付款的身份驗證情境 (1) |
✅ 1/1 (100.0%) |
✅ 1/1 (100.0%) |
❌ 0/1 (0.0%) |
❌ 0/1 (0.0%) |
| 用戶要求開立發票情境 (2) |
⚪ 1/2 (50.0%) |
⚪ 1/2 (50.0%) |
⚪ 1/2 (50.0%) |
⚪ 1/2 (50.0%) |
| 知識與產品查詢 (3) |
✅ 3/3 (100.0%) |
✅ 3/3 (100.0%) |
❌ 0/3 (0.0%) |
❌ 0/3 (0.0%) |
| 限時表演 / 指令 n8n 自動化退鑽（auto）勿關！ (1) |
❌ 0/1 (0.0%) |
❌ 0/1 (0.0%) |
❌ 0/1 (0.0%) |
❌ 0/1 (0.0%) |
Drop-off 最大的 5 個 scenario
- early_return drop at step1 (-100.0pp)
- 用戶申請開通超商代碼付款、轉帳付款的身份驗證情境 drop at step3 (-100.0pp)
- 知識與產品查詢 drop at step3 (-100.0pp)
- 限時表演 / 指令 n8n 自動化退鑽（auto）勿關！ drop at step1 (-100.0pp)
- 用戶問題回報情境 drop at step1 (-66.7pp)
Audit
Reproduce combined report: bin/rails runner "puts Eval::EvaluationReport.call(run: EvalRun.find(209))"
Or fetch each item separately:
bin/rails runner "puts Eval::KbAccuracyReport.call(run: EvalRun.find(209))"bin/rails runner "puts Eval::ScenarioFunnelReport.call(run: EvalRun.find(209))"
Per-Scenario × Per-Dim — Run #209
Suite: SWAG AI Bot (bulk R1) · scenarios: 7 · dims: 5 · populated cells: 32/35
| Scenario |
Scenario |
Tool |
Retrieval |
Faith |
AnsQ |
| early_return |
0.0% [0.0–0.0] (n=9) |
0.0% (n=1) |
100.0% [100.0–100.0] (n=8) |
100.0% [100.0–100.0] (n=9) |
88.9% [75.2–97.4] (n=9) |
| 用戶反映付款/儲值後未收到鑽石 |
100.0% (n=1) |
100.0% (n=1) |
0.0% (n=1) |
0.0% (n=1) |
100.0% (n=1) |
| 用戶問題回報情境 |
33.3% [0.0–100.0] (n=3) |
— |
0.0% [0.0–0.0] (n=3) |
44.4% [0.0–66.7] (n=3) |
42.2% [26.7–66.7] (n=3) |
| 用戶申請開通超商代碼付款、轉帳付款的身份驗證情境 |
100.0% (n=1) |
— |
0.0% (n=1) |
0.0% (n=1) |
46.7% (n=1) |
| 用戶要求開立發票情境 |
100.0% [100.0–100.0] (n=2) |
0.0% [0.0–0.0] (n=2) |
— |
83.3% [66.7–100.0] (n=2) |
68.3% [36.7–100.0] (n=2) |
| 知識與產品查詢 |
100.0% [100.0–100.0] (n=13) |
100.0% [100.0–100.0] (n=13) |
80.0% [50.0–100.0] (n=10) |
97.4% [92.3–100.0] (n=13) |
75.9% [61.0–89.2] (n=13) |
| 限時表演 / 指令 n8n 自動化退鑽（auto）勿關！ |
100.0% (n=1) |
0.0% (n=1) |
0.0% (n=1) |
100.0% (n=1) |
100.0% (n=1) |
Worst-3 cells (lowest primary score)
- 限時表演 / 指令 n8n 自動化退鑽（auto）勿關！ × Retrieval · 0.0% (n=1) · lowest sub_metric:
knowledges_source_recall
- 限時表演 / 指令 n8n 自動化退鑽（auto）勿關！ × Tool · 0.0% (n=1) · lowest sub_metric:
tools_recall
- 用戶要求開立發票情境 × Tool · 0.0% (n=2) · lowest sub_metric:
tools_recall