llm_eval_dispatch — Orchestrator

Graph

dispatch (llm_eval_dispatch)

Node Configuration

dispatch (llm_eval_dispatch)

Key	Value
`llm`
`model`
`temperature`

Saves to disk and updates any active crons using this pipeline

Run with Parameters

Pipeline JSON

{
  "edges": [
    {
      "to": "dispatch",
      "from": "__start__"
    },
    {
      "to": "__end__",
      "from": "dispatch"
    }
  ],
  "nodes": [
    {
      "name": "dispatch",
      "type": "llm_eval_dispatch",
      "config": {
        "llm": {
          "model": "gpt-4o",
          "temperature": 0
        }
      }
    }
  ],
  "description": "Online LLM eval: load emit artifacts, classify purposes (cheap rules or LLM), fan-out checklist judges, persist to DDB/S3. Triggered by ace-events llm_eval_trigger with runtime_params from llm.complete."
}