Share
ReportI am working on a prompt to summarize legal contracts in AI Builder GPT, but I am facing two issues:
-
Prompt filtering error: The token is being flagged due to some legal terms. As a workaround, I managed to bypass this by translating the text into English via Python before sending it to AI Builder. This way, it processes correctly and returns the expected output in Portuguese. However, this is not the best solution.
-
Token limitation: AI Builder seems to have a limit of 128k tokens, while in the GPT Playground on Azure, I can set it up to 150M tokens. This directly impacts the size of texts I can process.
In Azure, there is also an option to adjust the content filtering level, which I would set to "Low" to avoid this issue.
Is there any way to modify the filtering level in AI Builder? And is there any alternative to increase the token limit?
Is there any forecast for a future feature that includes these two configurations?
I appreciate any help!
All responses (
Answers (
519
