Skip to main content
2024-12-05

Copilot Enhancements

Span Chat

The Copilot Span Chat skill makes getting value from spans faster and easier. Rather than spending time scrolling through and deciphering span data , teams can now:
  • Analyze spans to extract key insights
  • Ask questions to quickly understand span data
  • Run evaluations on individual spans

Span Chat Evaluation

Dashboard Widget GeneratorBuilding dashboard plots just got way easier. The dashboard skill lets teams:
  • Create time series plots or distributions from natural language
  • Translate code (like Plotly) into ready-to-go visualizations
  • Handle ambiguous filters like “west coast states” and plot multiple widgets at once

Dashboard generator

Misc. Copilot Updates
  • We’ve revamped the main chat experience to be always accessible on the page, with an option to collapse the input bar
  • The Custom Metric skill now supports a conversational flow, making it easier for users to iterate and refine metrics dynamically

Additional Enhancements

Experiment Projects

Experiment traces for a dataset are now consolidated and can be accessed under “Experiment Projects” on the “Projects & Models” page.

Experiment Projects

Multi-Class/Label Per-Class Calibration & Chart

We’ve just rolled out per-class calibration metrics and calibration chart. Users can see calibration scores for each class separately and view the calibration chart all in one place.
  • To view per-class calibration simply select calibration from the metric dropdown and choose a class

Per-class calibration

  • The calibration chart can be found under the “More Charts” tab

Calibration Chart

SDK Version 7.29.0

  • Log experiments from a previously created dataframe

📚 New Content

The latest video tutorials, paper readings, ebooks, self-guided learning modules, and technical posts:🧑‍⚖️ Agent-as-a-Judge: Evaluate Agents with Agents🤖 LLM-as-a-Judge Evaluation for GenAI Use-Cases🌎 Building an AI Agent that Thrives in the Real World🛠️ AI Agent Workflows and Architectures Masterclass🔬 Agents in the Wild: Geotab