Evaluations are crucial for understanding and improving the quality of your AI prompt responses. Latitude provides a comprehensive evaluation framework to assess performance against various criteria.

​
Why Evaluate Prompts?

  • Measure Quality: Objectively assess if prompts meet desired standards (accuracy, relevance, tone, safety, etc.).
  • Identify Weaknesses: Pinpoint scenarios where prompts underperform.
  • Compare Versions: Quantify the impact of prompt changes (A/B testing).
  • Drive Improvement: Gather data to refine prompts using Prompt Suggestions.
  • Ensure Reliability: Build confidence in production-deployed prompts.

​
Evaluation Types in Latitude

Latitude supports three main approaches to evaluation, each suited for different needs:
  1. LLM-as-Judge:
    • How it works: Uses another language model (the “judge”) to score or critique the output of your target prompt based on specific criteria (e.g., helpfulness, clarity, adherence to instructions).
    • Best for: Subjective criteria, complex assessments, evaluating nuanced qualities like creativity or tone.
    • Requires: Defining evaluation criteria (often via templates or custom instructions for the judge LLM).
  2. Programmatic Rules:
    • How it works: Applies code-based rules and metrics to check outputs against objective criteria.
    • Best for: Objective checks, ground truth comparisons (using datasets), format validation (JSON, regex), safety checks (keyword detection), length constraints.
    • Requires: Defining specific rules (e.g., exact match, contains keyword, JSON schema validation) and potentially providing a Dataset with expected outputs.
  3. Human-in-the-Loop (Manual Evaluations):
    • How it works: Team members manually review prompt outputs (logs) and assign scores or labels based on their judgment.
    • Best for: Capturing nuanced human preferences, evaluating criteria difficult for LLMs to judge, initial quality assessment, creating golden datasets for other evaluation types.
    • Requires: Setting up manual review workflows and criteria for reviewers.

​
How Evaluations Connect to Prompts

  • Per-Prompt Basis: Evaluations are configured individually for each prompt within a project.
  • Target Logs: Evaluations run on the Logs generated by their associated prompt.
  • Triggering: Evaluations can be run manually on batches of logs/datasets or automatically on incoming logs (live mode). See Running Evaluations.
  • Results: Evaluation results (scores, labels, feedback) are stored alongside the corresponding logs, providing a rich dataset for analysis and improvement.

​
Actual Outputs

The actual output is the generated output from the model conversation. This is the output to perform evaluations against.

​
Selecting the Actual Output to evaluate against

By default the actual output is the last assistant message in the conversation, parsed as a simple string. However, some use cases requires more complex parsing, like evaluating tool calling or middle CoT.
  1. Go to the evaluation’s settings by clicking on the right-side button in the evaluation’s dashboard: Evaluation settings button
  2. Click on Advanced configuration.
  3. Configure:
    • Message selection: The last message or all messages in the conversation.
    • Content filter: Filter the messages by content type (e.g., text, images, tool calls…).
    • Parsing format: The format to parse the actual output into (e.g., string, JSON, YAML…).
    • Field accessor: The field to access in the actual output (e.g., ‘answer’, ‘arguments.recommendations[2]’…). Evaluation actual output configuration
  4. Test the configuration by clicking on the “Test” button: Evaluation actual output test

​
Expected Outputs

The expected output, also known as label, refers to the correct or ideal response that the language model should generate for a given prompt. You can create Datasets with Expected Output Columns to evaluate prompts with ground truth. Evaluation expected output

​
Negative Evaluations

Sometimes, you want to measure undesirable traits (e.g., toxicity, hallucination presence), where a lower score is better. Latitude allows you to mark evaluations as “negative”.
  1. Go to the evaluation’s settings by clicking on the right-side button in the evaluation’s dashboard: Evaluation settings button
  2. Click on Advanced configuration.
  3. Select “optimize for a lower score” to indicate high scores are undesirable: Evaluation negative setting
The Prompt Suggestions feature will use this setting to optimize correctly.

​
Next Steps

Dive deeper into each evaluation type: