[feature request] Base64 images rendered by default in Dataset (overview) + Evaluation (comparison) view

[TeoZosa]

Hey Wandb team, thanks for the hard work. Loving the feature set and dev velocity on the project so far.

Quick question: auto-rendering of base64 images has been super helpful for Traces¹, but it's conspicuously absent in the Datasets(overview)² & Evaluations comparison components. Are there any plans to add that feature to those views as well?

Examples

Datasets (overview)

Evaluations (comparison)

Footnotes

1. IIUC, added in feat: render anthropic input images as thumbnails #1859 courtesy of @jamie-rasmussen ↩

2. To be fair, this is more of a consequence of not being able to expand multi-element collections. Drilling down into single examples works. ↩