72일차 정답 추가 (실습 정답본 · answers · VIBE-A · tbench)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
mcchaeandClaude Opus 5.5 committed 2026-10-26 20:15:51 +09:00
1 parent baf72561f7
commit d88c5122d5
105 files changed
+24165

No files matched your search

+92
View File
@@ -0,0 +1,92 @@
$ uv run tools/flow.py bench my-helpdesk requests/faq-vpn-nocache.json -n 5
POST http://localhost:5678/webhook/my-helpdesk × 5
✔ #1 · HTTP 200 · 2.57s · route=자동 답변, llm_calls=3, cached=false
✔ #2 · HTTP 200 · 3.08s · route=자동 답변, llm_calls=3, cached=false
✔ #3 · HTTP 200 · 2.91s · route=자동 답변, llm_calls=3, cached=false
✔ #4 · HTTP 200 · 2.99s · route=자동 답변, llm_calls=3, cached=false
✔ #5 · HTTP 200 · 2.80s · route=자동 답변, llm_calls=3, cached=false
p50 2.91s · p95 3.08s · max 3.08s · failures 0/5
report: runs/bench-my-helpdesk-20261024-174351.json
$ uv run tools/flow.py bench my-helpdesk requests/ticket-own.json -n 5
POST http://localhost:5678/webhook/my-helpdesk × 5
✔ #1 · HTTP 200 · 2.86s · route=자동 답변, llm_calls=3, cached=false
✔ #2 · HTTP 200 · 3.05s · route=자동 답변, llm_calls=3, cached=false
✔ #3 · HTTP 200 · 3.12s · route=자동 답변, llm_calls=3, cached=false
✔ #4 · HTTP 200 · 2.94s · route=자동 답변, llm_calls=3, cached=false
✔ #5 · HTTP 200 · 2.99s · route=자동 답변, llm_calls=3, cached=false
p50 2.99s · p95 3.12s · max 3.12s · failures 0/5
report: runs/bench-my-helpdesk-20261024-174406.json
$ uv run tools/flow.py bench my-helpdesk requests/off-topic.json -n 5
POST http://localhost:5678/webhook/my-helpdesk × 5
✔ #1 · HTTP 200 · 2.38s · route=자동 답변, llm_calls=3, cached=false
✔ #2 · HTTP 200 · 2.36s · route=자동 답변, llm_calls=3, cached=false
✔ #3 · HTTP 200 · 2.48s · route=자동 답변, llm_calls=3, cached=false
✔ #4 · HTTP 200 · 2.49s · route=자동 답변, llm_calls=3, cached=false
✔ #5 · HTTP 200 · 2.84s · route=자동 답변, llm_calls=3, cached=false
p50 2.48s · p95 2.84s · max 2.84s · failures 0/5
report: runs/bench-my-helpdesk-20261024-174419.json
$ uv run tools/flow.py call my-helpdesk requests/ticket-own.json -v
HTTP 200 · 2.8s
→ request POST /webhook/my-helpdesk
{
"question": "T-1002 처리 상태 알려 주세요",
"user": "박서연"
}
← response HTTP 200
{
"route": "자동 답변",
"category": "ticket",
"answer": "T-1002 티켓은 현재 인증서 발급 승인 대기 중이며, 10월 20일에 완료될 예정입니다. 추가 문의 사항이 있으시면 010-2718-2818로 연락해 주시기 바랍니다.",
"sources": [],
"llm_calls": 3,
"cached": false,
"steps": [
"classify",
"faq",
"ticket",
"answer",
"review"
],
"timings": {
"classify_ms": 818,
"data_ms": 43,
"answer_ms": 877,
"review_ms": 895,
"total_ms": 2647
}
}
$ uv run tools/flow.py call my-helpdesk requests/off-topic.json -v
HTTP 200 · 2.5s
→ request POST /webhook/my-helpdesk
{
"question": "오늘 점심 메뉴 추천해 주세요",
"user": "이준호"
}
← response HTTP 200
{
"route": "자동 답변",
"category": "other",
"answer": "찾으시는 정보를 확인하지 못했습니다. 점심 메뉴 추천은 담당 직원에게 문의해 주시기 바랍니다.",
"sources": [],
"llm_calls": 3,
"cached": false,
"steps": [
"classify",
"faq",
"ticket",
"answer",
"review"
],
"timings": {
"classify_ms": 732,
"data_ms": 19,
"answer_ms": 740,
"review_ms": 978,
"total_ms": 2482
}
}