AutoEval
Generate, validate, execute, compare, and judge.
Configure
Pipeline Run
Metrics
History
New evaluation run
Evaluation preset
Questions CSV
Evaluation mode
Start run
Optional settings
Advanced pipeline settings
Evaluation context
Team name
Team ID
Team gender
League
Execution settings
Pipeline execution
SQL batch
TinyBird concurrency
Scout batch
TinyBird retries
Retry backoff
Judge concurrency
Scout Agent environment
Environment
SQL generation
SQL Scout model
SQL temperature
Scout response
T2S Scout model
T2S temperature
Video response
Video agent model
Video temperature
Judge
Judge model
Judge max tokens
Judge temperature