Glossary · Meaning and specification
Evidence viewV
2InstrumentWhat the evaluator may observe: the final answer only, or also sources, tool traces, hidden state, snapshots, provenance and logs.
Defined in
Handbook · InstrumentAnchors and evidence viewsWhat the judgment is rooted in, and what the evaluator may see: two coordinates most taxonomies forget, and the cheapest lever for better evaluators.Where it comes up
- Vocabulary without category errorsHandbook · Foundations
- Specification: what counts as successHandbook · Instrument
- LLM judges in practiceHandbook · Instrument
- Harness architecture and reproducibilityHandbook · Lifecycle
- Evaluating agentsHandbook · Agents and self-improvement
- Reading agent benchmarksHandbook · Agents and self-improvement
- Verifiability engineeringHandbook · Practice
- D: Autonomous research agentHandbook · Case studies
- Notation and formula cardHandbook · Reference
- CrosswalkHandbook · Reference
- The frameworkPage