Anthropic is expanding the developer console with useful features for instant creation and evaluation. Additionally, so-called artifacts can now be shared and redesigned using the Remix function.
Anthropic, an AI company in which Google and Amazon have invested heavily, introduced three cutting-edge AI solutions with Claude Opus, Sonnet, and Haiku in March 2024 — and surpassed them again in June 2024 with the launch of Claude 3.5 Sonnet. According to Anthropic, the model is twice as fast as Claude 3 Opus, can understand complex instructions better, and even outperforms OpenAI’s GPT-4o in some tests. Based on the latest technology, the company now has New features for the developer console It is developed so that users can create, test, compare and evaluate claims.
We've added new features to the Human Console.
Claude can create claims, create test variables, and show the claim outputs side by side. pic.twitter.com/H6THdoja7Z
– Anthropic (@AnthropicAI) July 9, 2024
Create, compare and evaluate claims
The console now provides developers with a built-in prompt generator based on Claude 3.5 Sonnet. Here, users can enter a specific task for the tool (for example, “prioritize incoming customer inquiries”), and a high-quality prompt will be generated.
Developers can use Claude to create input variables for prompts and run them to see the tool's response. Alternatively, different test cases can be entered manually.
Use Claude to create input variables for your router. Then run the router to see Claude's response.
You can also enter variables manually. pic.twitter.com/pJRKMJjCQ2
– Anthropic (@AnthropicAI) July 9, 2024
Additionally, the AI company introduced a feature that allows users to directly compare the outputs of two or more prompts. The answers, which are displayed side by side, can be rated from one point (“poor”) to five points (“excellent”).
We've also added the ability to compare the outputs of two or more claims side by side.
As you iterate over different versions of the prompt, your subject matter experts can compare and rate answers on a 5-point scale. pic.twitter.com/FpHvgka6z5
– Anthropic (@AnthropicAI) July 9, 2024
Artifacts can now be shared.
The new “Evaluate” tab aims to compare generated claims with real-world inputs and allow developers to fine-tune their test cases if needed and execute them all with one click. Anthropy to explain:
Testing claims against a set of real-world inputs can help you build confidence in the quality of your router before deploying it to production. With the new evaluation feature, you can do this directly in our console instead of manually running tests via spreadsheets or code.
The new Evaluate tab lets you automatically create test cases to evaluate your claim against real-world inputs.
Modify your test cases as needed, then run them all with one click. pic.twitter.com/HnNzzaXuYh
– Anthropic (@AnthropicAI) July 9, 2024
In addition to the new features for the developer console, Anthropic also announced an update to Artifacts that was introduced in June 2024. In the Artifacts workspace, users can ask Claude to create content such as code snippets, text documents, or website designs. These are displayed in a separate window. Users have recently been able to publish and share artifacts created with Claude, the company announced on X. Additionally, other users’ artifacts can be redesigned using the Remix feature.
“Certified tv guru. Reader. Professional writer. Avid introvert. Extreme pop culture buff.”
More Stories
Pitch: €56m for energy startup Reverion
Plastoplan: Plastics for Energy Transition
Canon Launches Arizona 1300 Series with FLXflow Technology