To validate an LLM’s performance, navigate to the Run Task view. You can access this view either from your home page or immediately after creating a new task.

In the Run Task view, the left panel displays your task configuration—including the user prompt, system prompt, chosen model, and defined outputs.

You can then enter or upload the input data for the task. Remember, the task is set up to run on dynamic input variables (e.g., {customer_review}). In other words, you instruct the LLM to execute the prompts on the provided input and produce outputs in the required format.

🧠 Dynamic inputs can be text, images or a combination of both. Each task can have as many dynamic input variables as required. For example, you might you might have a task that compares text to images, or summarises a range of numerical and text data.

For this demo, enter the following dummy {customer_review} in the input field:

"My toaster exploded during breakfast, sending flaming bread across the kitchen! 😱 On the bright side, I've discovered a new way to heat up the whole house. But seriously folks, this isn't just a hot topic - it's a fire hazard! The warranty card didn't mention anything about impromptu fireworks displays. 🎆"

Click Run Task (or press ⌘↩️) to execute the task. Within seconds, the model’s response appears in the right panel.

Understanding the JSON Response

Here’s an example of a task run response:

{
  "task_id": "0193e350-0c5d-8681-9284-3ccac209d5fa",
  "task_revision_id": "0193e350-0e1e-7873-a437-d1be2a9cae0a",
  "response": {
    "severity_level": 5,
    "critical_keywords": [
      "exploded",
      "flaming",
      "fire hazard"
    ],
    "immediate_actions": "Immediately contact the customer to ensure their safety and gather more details about the incident. Issue an urgent product recall for the specific toaster model. Initiate an internal investigation into the cause of the explosion. Prepare a public safety announcement warning other customers about the potential danger.",
    "overall_sentiment": "negative",
    "escalation_pathway": "Escalate immediately to Product Safety Team, Legal Department, and Executive Management. Notify relevant consumer safety regulatory bodies."
  },
  "run_data": {
    "submitted": {
      "customer_review": "My toaster exploded during breakfast, sending flaming bread across the kitchen! 😱 On the bright side, I've discovered a new way to heat up the whole house. But seriously folks, this isn't just a hot topic - it's a fire hazard! The warranty card didn't mention anything about impromptu fireworks displays. 🎆"
    },
    "files": []
  },
  "id": "0194d6de-6c37-7e27-4d9b-7a90092c858b",
  "created": "2025-02-05T16:07:57.528221",
  "input_tokens": 834,
  "output_tokens": 153,
  "total_tokens": 987,
  "input_processor_timing": 0.000009459967259317636,
  "llm_call_timing": 1.064588805020321
}

Key Points in the Response:

• Task ID & Revision ID: Unique identifiers required when calling the task API.

• Response Section: Contains the outputs in the specified format (e.g., sentiment, keywords, actions).

• Run Data: Displays the input data and any files submitted.

• Metadata: Includes a unique run ID, timestamp (created), token counts (input, output, total), and processing times for troubleshooting and performance assessment.

Iterating on Your Task: Based on the results, refine your prompts, experiment with different models, adjust parameters like temperature, or modify your output structures. You can also test the task with various input data to optimize performance.

Bonus Feature: Comparative Evaluation If you have a target output in mind, use the Rightbrain assistant’s evaluation feature. This tool compares your task output against an ideal response and provides suggestions for improvement.

From here, you can see an improved prompt or jump straight into Compare mode to compare your new prompt against your old one.