Our product team has been busy this week deploying significant developments to our new explorer anomaly detection framework. Here are some of the updates we’ve shipped to our front end. We thank you for your continued support and collaboration, since it directly helps us continue to ship great products for our customers.

Product Updates

Anomaly detection events now appear on the Radiant Home page

Radiant Explorer provides more than a dozen metrics and evaluations of generative AI traffic. Creating filters allows users to detect anomalies faster by defining multiple criteria that could represent outliers. Now users can access their saved filters and surface anomalies even faster from the Radiant home page. Radiant customers can consume filter responses in the platform or through their own observability tools.

New evaluation options to detect anomalies

This week we added a suite of additional evaluations as part of our out of the bod configuration. These evaluators are designed to help catch sensitive content being returned by LLMs including:

• Tone analysis

• Conciseness

• Refusal to answer

• Helpfulness

• Bias (Age, Race, Gender)

Create Custom Metrics

Users in explorer now have the ability to create their own custom metrics, defining the specific cases and criteria for these metrics. The Radiant platform can compute results for these metrics based on historical data and categories future logs in real time.

Improved filters UI and expanded filter options

This week we made it easier to add and save filters by allowing users to select all filters from the filter list. This allows users to quickly select multiple filters and customize them so that they can drill down on the dashboard or logs view. We also added a new filter based on the access group related to the prompt-response pair.

About Radiant

Radiant is the Enterprise AI platform to take you from idea to production deployment. From security and governance to scaling and resiliency, we make it simple to make AI a critical part of your business. 

Try out a demo here, sign up here to get your own instance, or reach out to our founders directly at [email protected].

We’re also hiring. If you know of someone great that is interested in helping every company build AI into their products and operations, we'd love an introduction.