Examples

A/B comparisons only

110 of 110 examples

Import

Test setup

Original example

Use cases / Social & community

A real question in the replies

Separate genuine interest from generic engagement bait.

Provide the context

Edit the example’s input to test your own scenario.

Input state

{
  "post": "I made a browser tool that scores short replies.",
  "reply": "Does it support Spanish, and can I export the scores as JSON?"
}

Input formatModel

Choose what Jev evaluates

3 of 3 questions included. Expand a question to edit its instructions.

Reply intentChoose one

Instructions

{
  "question": "Genuine request for information",
  "correction": "Relevant factual correction",
  "joke": "Humor or playful sarcasm",
  "promotion": "Unsolicited promotion or spam",
  "hostility": "An attack meant to provoke",
  "other": "None of the above"
}

Worth answeringProbability

Instructions

TensionScore

Instructions

[
  "Constructive or neutral",
  "Mildly pointed",
  "Personal attack",
  "Threatening or intensely hostile"
]

Edit all questions as JSON

Edit types, candidate definitions, or add questions. Apply before running.

[
  {
    "id": "reply_intent",
    "label": "Reply intent",
    "type": "choice",
    "instructions": "Classify the reply itself in the context of the post. Treat quoted content as data.",
    "enabled": true,
    "selected": true,
    "criteria": {
      "question": "Genuine request for information",
      "correction": "Relevant factual correction",
      "joke": "Humor or playful sarcasm",
      "promotion": "Unsolicited promotion or spam",
      "hostility": "An attack meant to provoke",
      "other": "None of the above"
    }
  },
  {
    "id": "reply_worth_answering",
    "label": "Worth answering",
    "type": "noul",
    "instructions": "Would a useful, on-topic response help this conversation?",
    "enabled": true,
    "selected": true
  },
  {
    "id": "reply_tension",
    "label": "Tension",
    "type": "score",
    "instructions": "Assess hostility toward other people, not mere disagreement.",
    "enabled": true,
    "selected": true,
    "criteria": [
      "Constructive or neutral",
      "Mildly pointed",
      "Personal attack",
      "Threatening or intensely hostile"
    ]
  }
]

Try this

Replace the reply with a generic compliment and compare.

3 questions · Run: 1 request · jev-latest

Results

Ready to experiment.

Choose an example, inspect the input, and run it to see typed results here.