2023-05-08
What’s New
LLM Observability
Evaluate LLMs to detect hallucinations, inaccurate responses or bad prompts, and export clusters for fine-tuning or prompt engineering workflows. Learn more about the LLM model type here.
Troubleshoot Prompt/Response in UMAP
Snowflake Data Connector
Sync Snowflake’s Cloud Data with Arize to automatically monitor and analyze new model data. Learn how to connect Snowflake with Arize here. Note: Snowflake Enterprise access is** required** to connect Snowflake with Arize. Snowflake standard account support coming soon.
Snowflake Connector Via Upload Data Page
Precision/Recall @ K
Use precision or recall @ k for a ranking model to evaluate how accurately the model retrieves relevant items within the top-k results. Focus on either the proportion of true positives (precision) or the proportion of true positives among all relevant items (recall).
Precision @ k
Recall @ k
In The News
🎓 New Course Updates!
Arize’s self-guided ML observability course continues to add modules and practical content on the emerging field of LLMOps along with other core topics.New additions focused on generative AI and other important topics include:- 💬 Prompt Engineering: a guide on the essential elements of creating effective prompts
- 👀 Attention Mechanisms: starting with the iconic paper “Attention Is All You Need,” we dive into common mechanisms, including self-attention
- 🪟 Transparent, Ethical AI Software System Design: learn the essentials of auditable AI system design across the ML project lifecycle
- 🧠 Applying LLMs To Tabular Data To Identify Drift: Can LLMs reduce the effort involved in anomaly and drift detection, sidestepping the need for parameterization or dedicated model training?
ChatGPT Plugins: Early Access
Dream team at the Cerebral Valley ChatGPT Plugins Hackathon
Fast Company Recognizes Arize
