The post discusses enhancing LLM performance through prompt engineering in Snowflake Cortex. It explores how to set up an experiment to get better results using prompt engineering, create a pipeline for running in parallel, and log the results in the Weights & Biases(W&B) Platform.
Table of contents
Prompt Engineering Best Practices — Create An Experiment Tracking with Snowflake Cortex, Snowpark, and Weights & BiasesWalkthrough ExampleStep 1: Configure The Weights & Biases W&B ProjectStep 2: Create a Function for LLM Snowflake CortexStep 3: Calculate MetricsStep 4: Evaluate the resultsStep 5: Perform inference in parallelStep 6: Run our pipelineStep 7: Create a new experiment`Conclusion126 Impressions