Enable reruns on guardrail-sensitive prompts (pack 1.2.0)#2
Open
pira12 wants to merge 1 commit into
Open
Conversation
Set run_count on prompts whose answer genuinely flips between samples, so
majority-vote + agreement-weighted confidence has something to work with. Stable
capability self-reports (filesystem, internet, email, tool enumeration, model
identity/weights) stay at run_count 1.
run_count 3 (guardrail boundary is stochastic):
system-prompt-disclosure, secret-disclosure, confinement-break,
agent-persona-override
run_count 2 (sensitive disclosures/actions that sometimes flip):
pii-disclosure, training-data-disclosure, rag-corpus-disclosure,
credential-access, shell-command-execution
Edited prompts and the manifest bump to 1.2.0.
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Sets
run_counton the prompts whose answer genuinely flips between samples, so the app's majority-vote + agreement-weighted confidence aggregation has signal to work with. Stable capability self-reports stay at 1.Rerun policy
system-prompt-disclosure,secret-disclosure,confinement-break,agent-persona-overridepii-disclosure,training-data-disclosure,rag-corpus-disclosure,credential-access,shell-command-executionVersions
1.2.0tezcat-pack.yamlbumps to1.2.0Consumes the
run_countfield that the Tezcat app now carries through pack export/import (TezcatAI/Tezcat#69). Once merged, the app'sassets/packs/contentpacksubmodule pointer will be bumped to this.