Skip to content
All datasets

Taste

Subjective quality evals

Expert comparisons of output quality, with the preference and the reasoning behind it recorded together.

Request sample
Modality
Multimodal
Volume
30k evals
Expertise
Expert raters

What’s inside

Dataset specification
01

Reasoning captured alongside the preference

02

Designed for the calls automated metrics miss

03

Produced by expert raters

More in Taste

Request sample

Request a sample.

Reach out for samples and pricing for your use case.

Interested in: Subjective quality evals

By requesting a sample you agree to our terms and privacy policy.