benchclaw-stage4-invalid-item-screening

To filter out invalid items from grey-batch validation data, run the screening scripts to identify media missing, answer missing, option errors, and scoring contract issues, ensuring only valid items are processed for model evaluation.

Item screening detects data integrity issues including missing media, missing answers, incorrect options, missing scoring contracts, and evidence traceability problems before model evaluation or full-scale synthesis begins.

Grey-batch validation fails because missing media or answer fields break data integrity, preventing accurate model evaluation and requiring invalid item screening to remove incomplete data points before processing.

You need to run invalid item screening before conducting model evaluations or full-scale synthesis to ensure data integrity and prevent invalid items from affecting the accuracy and reliability of results.

Item screening requires no external dependencies and operates using included scripts to process grey-batch validation data, identifying and filtering invalid items for subsequent model evaluation workflows.

The best way to ensure data quality in grey-batch validation is to run invalid item screening scripts that detect and filter out items with missing media, answers, option errors, and scoring contract issues before full-scale synthesis.