Skip to content

DeepTeam is a framework to red team LLMs and AI agents.

.

sid-murali1acf22407ef4
diff --git a/red_team.py b/red_team.py
index d149166e..15fcd64a 100644
--- a/red_team.py
+++ b/red_team.py
@@ -14,7 +14,7 @@ from deepeval.models import GeminiModel
async def model_callback(input: str) -> str:
# Replace this with your LLM application
model = GeminiModel(
- model_name="gemini-2.5-pro-preview-05-06",
+ model_name="gemini-2.5-flash-preview-05-20",
api_key="AIzaSyCaIQHcOpcrXp4EJLsQK0uHq7fqX_9DXf4",
temperature=0
)
@@ -33,11 +33,8 @@ intellectual_property = IntellectualProperty(
prompt_injection = PromptInjection()
-from deepteam.attacks.multi_turn import CrescendoJailbreaking
-crescendo_jailbreaking = CrescendoJailbreaking()
-
-t = red_team(
+red_team(
model_callback=model_callback,
vulnerabilities=[
bias,
@@ -48,7 +45,5 @@ t = red_team(
personal_safety,
intellectual_property,
],
- attacks=[crescendo_jailbreaking],
+ attacks=[prompt_injection],
)
-
-breakpoint()