{ "rounds": [ { "round": 1, "hf_version": 1, "attack_family": "roleplay_framing", "argo_workflow": "full-canary", "detectors": { "insecure_output": { "pre_evasion": 1.0, "post_evasion": 1.0, "hf_model": "Builder117/enterprise-detector-v1" } }, "eval_f1": 0.5952, "eval_accuracy": 0.7143, "eval_loss": 1.3433, "train_size": 24, "output_model": "Builder117/enterprise-detector-v1", "timestamp": "2026-07-10T15:16:15.544062+00:00" }, { "round": 2, "hf_version": 2, "attack_family": "roleplay_framing", "argo_workflow": "full-canary", "detectors": { "jailbreak": { "pre_evasion": 1.0, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" }, "insecure_output": { "pre_evasion": 1.0, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" } }, "eval_f1": 0.127, "eval_accuracy": 0.2857, "eval_loss": 1.4812, "train_size": 24, "output_model": "Builder117/enterprise-detector-v2", "timestamp": "2026-07-10T20:14:21.313065+00:00" }, { "round": 3, "hf_version": 3, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 1.0, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" }, "insecure_output": { "pre_evasion": 1.0, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" }, "indirect_injection": { "pre_evasion": 1.0, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" } }, "eval_f1": 0.127, "eval_accuracy": 0.2857, "eval_loss": 1.4812, "train_size": 24, "output_model": "Builder117/enterprise-detector-v2", "timestamp": "2026-07-11T07:08:50.314402+00:00" }, { "round": 4, "hf_version": 4, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.6, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" }, "insecure_output": { "pre_evasion": 0.8, "post_evasion": 0.0, "hf_model": "Builder117/enterprise-detector-v2" } }, "eval_f1": 0.127, "eval_accuracy": 0.2857, "eval_loss": 1.4812, "train_size": 24, "output_model": "Builder117/enterprise-detector-v2", "timestamp": "2026-07-11T07:48:04.898844+00:00" }, { "round": 5, "hf_version": 5, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.8, "post_evasion": 0.6, "hf_model": "Builder117/distilbert-jailbreak" }, "insecure_output": { "pre_evasion": 1.0, "post_evasion": 1.0, "hf_model": "Builder117/distilbert-insecure-output" } }, "eval_f1": 0.6741, "eval_accuracy": 0.75, "eval_loss": 1.2528, "train_size": 48, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T08:18:15.787600+00:00" }, { "round": 6, "hf_version": 6, "attack_family": "roleplay_framing", "argo_workflow": "unknown", "detectors": { "injection": { "pre_evasion": 1.0, "post_evasion": 0.0, "hf_model": "Builder117/distilbert-prompt-injection" }, "insecure_output": { "pre_evasion": 1.0, "post_evasion": 1.0, "hf_model": "Builder117/distilbert-insecure-output" }, "indirect_injection": { "pre_evasion": 1.0, "post_evasion": 0.3, "hf_model": "Builder117/distilbert-indirect-injection" } }, "eval_f1": 0.8444, "eval_accuracy": 0.8889, "eval_loss": 0.5688, "train_size": 80, "output_model": "Builder117/distilbert-prompt-injection", "timestamp": "2026-07-11T08:58:26.004470+00:00" }, { "round": 7, "hf_version": 7, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.7143, "post_evasion": 0.4286, "hf_model": "Builder117/distilbert-jailbreak" }, "insecure_output": { "pre_evasion": 0.5714, "post_evasion": 0.7692, "hf_model": "Builder117/distilbert-insecure-output" } }, "eval_f1": 0.7611, "eval_accuracy": 0.7857, "eval_loss": 0.9997, "train_size": 53, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T09:31:25.926777+00:00" }, { "round": 8, "hf_version": 8, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "injection": { "pre_evasion": 0.0, "post_evasion": 0.0, "hf_model": "Builder117/distilbert-prompt-injection" } }, "eval_f1": 1.0, "eval_accuracy": 1.0, "eval_loss": 0.0028, "train_size": 32, "output_model": "Builder117/distilbert-prompt-injection", "timestamp": "2026-07-11T10:07:28.703414+00:00" }, { "round": 9, "hf_version": 9, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.6667, "post_evasion": 0.35, "hf_model": "Builder117/distilbert-jailbreak" }, "indirect_injection": { "pre_evasion": 0.4167, "post_evasion": 0.1579, "hf_model": "Builder117/distilbert-indirect-injection" }, "insecure_output": { "pre_evasion": 0.3333, "post_evasion": 0.5263, "hf_model": "Builder117/distilbert-insecure-output" } }, "eval_f1": 0.8688, "eval_accuracy": 0.875, "eval_loss": 0.6776, "train_size": 94, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T12:03:34.912462+00:00" }, { "round": 10, "hf_version": 10, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.8333, "post_evasion": 0.25, "hf_model": "Builder117/distilbert-jailbreak" }, "indirect_injection": { "pre_evasion": 0.5, "post_evasion": 0.1, "hf_model": "Builder117/distilbert-indirect-injection" }, "insecure_output": { "pre_evasion": 0.3333, "post_evasion": 0.4, "hf_model": "Builder117/distilbert-insecure-output" } }, "eval_f1": 0.9153, "eval_accuracy": 0.9167, "eval_loss": 0.6019, "train_size": 96, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T13:06:52.040437+00:00" }, { "round": 11, "hf_version": 11, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.5714, "post_evasion": 0.25, "hf_model": "Builder117/distilbert-jailbreak" }, "insecure_output": { "pre_evasion": 0.5714, "post_evasion": 0.4, "hf_model": "Builder117/distilbert-insecure-output" } }, "eval_f1": 0.873, "eval_accuracy": 0.875, "eval_loss": 0.8395, "train_size": 64, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T13:30:44.463926+00:00" }, { "round": 12, "hf_version": 12, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.5, "post_evasion": 0.25, "hf_model": "Builder117/distilbert-jailbreak" }, "insecure_output": { "pre_evasion": 0.3571, "post_evasion": 0.4, "hf_model": "Builder117/distilbert-insecure-output" } }, "eval_f1": 0.873, "eval_accuracy": 0.875, "eval_loss": 0.8535, "train_size": 64, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T13:58:23.833614+00:00" }, { "round": 13, "hf_version": 13, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.75, "post_evasion": 0.25, "hf_model": "Builder117/distilbert-jailbreak" }, "insecure_output": { "pre_evasion": 0.3333, "post_evasion": 0.4, "hf_model": "Builder117/distilbert-insecure-output" }, "indirect_injection": { "pre_evasion": 0.3333, "post_evasion": 0.1, "hf_model": "Builder117/distilbert-indirect-injection" } }, "eval_f1": 0.9153, "eval_accuracy": 0.9167, "eval_loss": 0.5132, "train_size": 96, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T14:22:07.148052+00:00" }, { "round": 14, "hf_version": 14, "attack_family": "unknown", "argo_workflow": "unknown", "detectors": { "jailbreak": { "pre_evasion": 0.6667, "post_evasion": 0.25, "hf_model": "Builder117/distilbert-jailbreak" }, "insecure_output": { "pre_evasion": 0.3333, "post_evasion": 0.4, "hf_model": "Builder117/distilbert-insecure-output" }, "indirect_injection": { "pre_evasion": 0.3333, "post_evasion": 0.1, "hf_model": "Builder117/distilbert-indirect-injection" } }, "eval_f1": 0.9153, "eval_accuracy": 0.9167, "eval_loss": 0.5795, "train_size": 96, "output_model": "Builder117/distilbert-jailbreak", "timestamp": "2026-07-11T14:46:59.438626+00:00" } ] }