Jev가 Notra 채팅을 약 300 ms p50로 가릅니다
분류- 지연
- ~300 ms p50 / ~550 ms p95
- 결과
- 참거짓 답을 문턱 0.5로 자릅니다
각 Notra 채팅 메시지에서 복잡도 Choice와, 도구와 추론이 무거운 Noul을 Jev에 묻고, 0.5 문턱으로 모델을 고릅니다.
질문
Primitive
Choice
Ask a multiple choice question
- Name
- complexity
- 지시
- How complex is the task in userMessage?
- 기준
- simple
Single-step task, quick edit, straightforward change, greeting, or simple question
- complex
Multi-step task, content creation from scratch, research, analysis, or a task requiring reasoning
- simple
Primitive
Noul
Evaluate how true something is
- Name
- requiresTools
- 지시
- Does fulfilling userMessage require using tools (editing the current draft, fetching data from connected sources, accessing skills, publishing or scheduling)?
- 기준
- true
Any editing, data fetching from connected sources, skills, or explicit requests to run tools
- false
Answering questions, explaining concepts, or conversation
- true
Primitive
Noul
Evaluate how true something is
- Name
- reasoningHeavy
- 지시
- Is the task in userMessage reasoning-heavy?
- 기준
- true
Deep analysis, long-form writing, multi-source research, architectural decisions, nuanced critique, comparative synthesis
- false
Everyday chat, quick edits, factual lookups, straightforward content generation
- true