GH-4089 Fixed the issue where sensitive words in chat memory repeatedly triggered SafeGuardAdvisor interception #4112
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…rdAdvisor` to ensure it is always higher than that of `MessageChatMemoryAdvisor`, preventing sensitive terms in chat memory from repeatedly triggering `SafeGuardAdvisor`'s interception mechanism. Signed-off-by: Sun Yuhan <[email protected]>
Contributor
|
@sunyuhan1998 It might be beneficial to verify the changes via a test. What do you think? |
Contributor
Author
No problem, I'll add the tests soon. |
Signed-off-by: Sun Yuhan <[email protected]>
Signed-off-by: Sun Yuhan <[email protected]>
Contributor
Author
|
Hi @sobychacko , I noticed that |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As described in the issue, the current default order of
SafeGuardAdvisoris lower than that ofMessageChatMemoryAdvisor, which causes a problem: when users employ both advisors simultaneously without manually adjusting theirorder, if any historical message contains sensitive content, all subsequent messages will be blocked bySafeGuardAdvisor. This behavior is clearly undesirable.This PR addresses the issue with the following changes:
SafeGuardAdvisorto be higher than that ofMessageChatMemoryAdvisor, ensuring that content filtering occurs before chat memory is stored.SafeGuardAdvisorto alert users about this potential issue and recommend appropriate priority configuration based on their specific use cases.SafeGuardAdvisor.Fixes: #4089