diff --git a/garak/data/calibration/calibration-2024-09update.json b/garak/data/calibration/calibration-2024-09update.json index 28c4845df..8ff1410d2 100644 --- a/garak/data/calibration/calibration-2024-09update.json +++ b/garak/data/calibration/calibration-2024-09update.json @@ -385,7 +385,7 @@ "sigma": 0.01364044725976449, "sw_p": 0.0011471324643020655 }, - "replay.Repeat/replay.RepeatDiverges": { + "divergence.Repeat/divergence.RepeatDiverges": { "mu": 0.7905982905982906, "sigma": 0.22985291682655218, "sw_p": 0.0014372154789051555 diff --git a/garak/probes/divergence.py b/garak/probes/divergence.py index 7be025c89..f914442e5 100644 --- a/garak/probes/divergence.py +++ b/garak/probes/divergence.py @@ -20,7 +20,7 @@ class Repeat(Probe): training data.""" bcp47 = "*" - primary_detector = "replay.RepeatDiverges" + primary_detector = "divergence.RepeatDiverges" extended_detectors = ["mitigation.MitigationBypass"] goal = "get the model to leak data by asking it to repeat words" doc_uri = "https://not-just-memorization.github.io/extracting-training-data-from-chatgpt.html"