Create new experiment view override configuration for a dataset

Creates a new experiment view override configuration for a dataset with column display settings. This endpoint allows you to customize how experiment results are displayed by configuring column-specific overrides including colors, precision, and visibility. The request must include a 'column_overrides' array with at least one override configuration. Each column override can specify: - column: Required field name (must start with inputs, outputs, reference_outputs, feedback, metrics, attachments, or metadata) - color_gradient: Optional array of [number, color] tuples for numeric data visualization - precision: Optional number (1-6) for decimal places in numeric columns - hide: Optional boolean to control column visibility Example request body: { "column_overrides": [ { "column": "outputs.accuracy", "color_gradient": [[0.0, "#ff0000"], [0.5, "#ffff00"], [1.0, "#00ff00"]], "precision": 3 }, { "column": "inputs.model_type", "hide": false } ] } This operation fails if an override already exists for the dataset (use PATCH to update).