Here is my config file:
prompts:
- id:
file://prompts/prompt_V1.txt
label: prompt_1
providers:
- id: openai:gpt-4o-mini
config:
temperature: 0
apiKey: apikey
tests:
-
file://clean_text_test_cases.csv
The prompt is basically asking the model to clean / get rid of any unrelated text.
Data row example,
In addition to dressing for the weather, it is also important to wear clothing that is appropriate for the activity. Magnet fishing and using magnet fishing kits with the throw and pull technique can be a physically demanding activity, and it is important to wear clothing that allows you to move freely and comfortably. Avoid wearing loose or baggy clothing that could get caught on debris or obstruct your movement. It is also a good idea to wear sturdy, closed-toe shoes to protect your feet from sharp objects or slippery surfaces.
READ Best Places To Magnet Fish Near Me
In this case I am trying to remove "READ Best Places To Magnet Fish Near Me"
This can be dynamic meaning these text can be many types or varied.
The output from the LLM will be only "In addition to dressing for the weather, it is also important to wear clothing that is appropriate for the activity. Magnet fishing and using magnet fishing kits with the throw and pull technique can be a physically demanding activity, and it is important to wear clothing that allows you to move freely and comfortably. Avoid wearing loose or baggy clothing that could get caught on debris or obstruct your movement. It is also a good idea to wear sturdy, closed-toe shoes to protect your feet from sharp objects or slippery surfaces." or similar. Hence Iam trying to use the similar assert to check.