07.17.2023
New Releases, Enhancements, Changes + Arize in the News!
What's New
LLMs: Open AI Cluster Summarization
With the overwhelming amount of text data processed by large language models (LLMs) every day, it's challenging to manually comb through each prompt or response for analysis and insights. Using an Open AI integration, Arize sends the raw text from your prompt or response embeddings to Open AI, and generates a summary - turning unwieldy data sets into easily understandable clusters. Learn more
Enhancements
Ranking: Prediction Score Drift
Visualize and inspect how ranking prediction scores drift over time to identify features within your model that negatively impact prediction relevancy.
This is only available for ranking models with prediction scores.
Dashboards: Line Chart Plots
Create multiple plots with different metrics in the same view for a simplified way to create and use line chart plots with dashboards.
Python SDK v7.1.0
Additional LLM and object detection model type support for single record logging.
Learn about Python SDK fixes and improvements here.
#creating LLM records for single record logging
prompt = Embedding(
vector= [0.10031, -0.06012, -0.00259, -0.08568, ...],
data="! Why does the UI have to be so clunky and unintuitive?! It's like every time I try to navigate through the options, I end up getting lost in a sea of confusing menus and poorly labeled buttons. And don't even get me started on the color scheme - who thought neon green on black was a good idea? At this point, I'm ready to throw my computer out the window and switch to a different product altogether. Come on, developers - make things simpler for us already!"
)
response = Embedding(
vector= [-0.11366, -0.20642, -0.03115, -0.18099,...],
data="I'm sorry to hear that you're experiencing difficulties with the UI. As an AI language model, I don't have the capability to change the UI, but I suggest looking into user guides, tutorials or support forums to help navigate through the menus more easily. Additionally, you could provide feedback to the developers on the UI, either through in-product feedback tools or their website. They may take into account your suggestions in future updates."
)
#creating object detection records for single record logging
object_detection_prediction = ObjectDetectionLabel(
bounding_boxes_coordinates: [[147.5, 98.0, 169.67, 276.69], [114.19, 108.4...]]
categories: ['chair', 'chair', 'bed', 'dining table', 'chair']
scores: [0.90, 0.97, 0.98, 0.96, 0.96, 0.92, ...]
)
object_detection_actual = ObjectDetectionLabel(
bounding_boxes_coordinates: [[437.51, 366.37, 529.0, 480.97], [0.0, 302.24...]]
categories: ['dining table', 'bed', 'chair', 'chair']
)In The News
📚 New AI Research Paper Readings
Catch up on the latest cutting-edge research in generative AI with these paper readings:
Orca Paper: Progressive Learning from Complex Explanation Traces of GPT-4
✈️ Arize Debuts Observe Copilot, Winning Coolest Technology at VB Transform's Innovation Showcase
Through an integration with GPT-4, Arize Observe Copilot's plugin for LLM observability provides an intuitive, chat-like experience for LLM practitioners to ask questions about their model, perform EDA, and uncover problematic clusters of data points.
🎉 Further Industry Recognition
Arize won several other awards this month:
💬 How To Troubleshoot LLM Summarization Tasks
This Phoenix code-along blog dives into how to find the root-cause of LLM performance issues for an LLM text summarization task by analyzing prompt-response pairs.
Interview: Mark Scarr, Senior Director of Data Science, Atlassian
Covering Atlassian's ML use cases and exciting plans around LLMs and generative AI.
Last updated
Was this helpful?