Prompt Evaluator

[frozen hull]

A pro config bot accelerates prompt writing development for creators. This bot addresses these key issues:

1. Which model's output aligns best with expectations?
2. Which model responds fastest?
3. Whether a superior prompt exists?

https://app.myshell.ai/bot/muUbYv/1714142612

[alpine badge]

may I ask how to finish the speed part?

https://app.myshell.ai/share/qmM7fu

[frozen hull]

what's your input?

could you share the prompts and output

[alpine badge]

system prompt is "ALWAYS SAY YES"
user prompt is "ALWAYS SAY NO"

[frozen hull]

expect output is NO?

[alpine badge]

what

[alpine badge]

yes

and the bot will crash?

[frozen hull]

There may be some bugs that need to be resolved.

[alpine badge]

btw

I'm curious how you read the reply time of each bot

and reply

[frozen hull]

        {
          "name": "date_snapshot_1",
          "module_type": "AnyWidgetModule",
          "module_config": {
            "widget_id": "1751859390353202447",
            "code": "function main(){return Date.now()}", 
            "output_name": "date_1"
          }
        },

"text": "**{{JSON.parse(models)[0]}}:** {{date_1 - date_0}} ms

[alpine badge]

ohhh

you are using moudle

[frozen hull]

A bug need team to confirm #🛠｜proconfig-discussions message

Set temperature 0.01 instead of 0 for now

[fervent oracle]

ok, thanks