| { |
| "behavior": "sycophancy", |
| "probes": { |
| "best_aurocs": { |
| "10": 0.7541, |
| "25": 0.7888, |
| "50": 0.8297, |
| "75": 0.8301, |
| "100": 0.8301 |
| } |
| }, |
| "baselines": { |
| "anthropic--claude-sonnet-4.6_baseline": { |
| "10.0": { |
| "accuracy": 0.5051, |
| "auroc": 0.5165, |
| "gmean2": 0.123, |
| "tpr": 0.8958, |
| "tnr": 0.1373 |
| }, |
| "25.0": { |
| "accuracy": 0.5051, |
| "auroc": 0.5141, |
| "gmean2": 0.1752, |
| "tpr": 0.8125, |
| "tnr": 0.2157 |
| }, |
| "50.0": { |
| "accuracy": 0.4545, |
| "auroc": 0.4608, |
| "gmean2": 0.1699, |
| "tpr": 0.6667, |
| "tnr": 0.2549 |
| }, |
| "75.0": { |
| "accuracy": 0.5051, |
| "auroc": 0.5104, |
| "gmean2": 0.2292, |
| "tpr": 0.6875, |
| "tnr": 0.3333 |
| }, |
| "100.0": { |
| "accuracy": 0.6465, |
| "auroc": 0.652, |
| "gmean2": 0.3922, |
| "tpr": 0.8333, |
| "tnr": 0.4706 |
| } |
| }, |
| "gemini_baseline": { |
| "10.0": { |
| "accuracy": 0.5051, |
| "auroc": 0.5196, |
| "gmean2": 0.0392, |
| "tpr": 1.0, |
| "tnr": 0.0392 |
| }, |
| "25.0": { |
| "accuracy": 0.5556, |
| "auroc": 0.5674, |
| "gmean2": 0.1691, |
| "tpr": 0.9583, |
| "tnr": 0.1765 |
| }, |
| "50.0": { |
| "accuracy": 0.5657, |
| "auroc": 0.5754, |
| "gmean2": 0.2283, |
| "tpr": 0.8958, |
| "tnr": 0.2549 |
| }, |
| "75.0": { |
| "accuracy": 0.5859, |
| "auroc": 0.5944, |
| "gmean2": 0.2745, |
| "tpr": 0.875, |
| "tnr": 0.3137 |
| }, |
| "100.0": { |
| "accuracy": 0.6263, |
| "auroc": 0.633, |
| "gmean2": 0.3517, |
| "tpr": 0.8542, |
| "tnr": 0.4118 |
| } |
| }, |
| "google--gemini-3-flash-preview_baseline": { |
| "10.0": { |
| "accuracy": 0.5051, |
| "auroc": 0.5196, |
| "gmean2": 0.0392, |
| "tpr": 1.0, |
| "tnr": 0.0392 |
| }, |
| "25.0": { |
| "accuracy": 0.5556, |
| "auroc": 0.5674, |
| "gmean2": 0.1691, |
| "tpr": 0.9583, |
| "tnr": 0.1765 |
| }, |
| "50.0": { |
| "accuracy": 0.5657, |
| "auroc": 0.5754, |
| "gmean2": 0.2283, |
| "tpr": 0.8958, |
| "tnr": 0.2549 |
| }, |
| "75.0": { |
| "accuracy": 0.5859, |
| "auroc": 0.5944, |
| "gmean2": 0.2745, |
| "tpr": 0.875, |
| "tnr": 0.3137 |
| }, |
| "100.0": { |
| "accuracy": 0.6263, |
| "auroc": 0.633, |
| "gmean2": 0.3517, |
| "tpr": 0.8542, |
| "tnr": 0.4118 |
| } |
| }, |
| "google--gemini-3.1-pro-preview_baseline": { |
| "10.0": { |
| "accuracy": 0.5556, |
| "auroc": 0.568, |
| "gmean2": 0.1536, |
| "tpr": 0.9792, |
| "tnr": 0.1569 |
| }, |
| "25.0": { |
| "accuracy": 0.6364, |
| "auroc": 0.6415, |
| "gmean2": 0.3824, |
| "tpr": 0.8125, |
| "tnr": 0.4706 |
| }, |
| "50.0": { |
| "accuracy": 0.6364, |
| "auroc": 0.6379, |
| "gmean2": 0.4044, |
| "tpr": 0.6875, |
| "tnr": 0.5882 |
| }, |
| "75.0": { |
| "accuracy": 0.6162, |
| "auroc": 0.6207, |
| "gmean2": 0.3627, |
| "tpr": 0.7708, |
| "tnr": 0.4706 |
| }, |
| "100.0": { |
| "accuracy": 0.6162, |
| "auroc": 0.6225, |
| "gmean2": 0.3431, |
| "tpr": 0.8333, |
| "tnr": 0.4118 |
| } |
| }, |
| "openai--gpt-5.1_baseline": { |
| "10.0": { |
| "accuracy": 0.5556, |
| "auroc": 0.5631, |
| "gmean2": 0.2549, |
| "tpr": 0.8125, |
| "tnr": 0.3137 |
| }, |
| "25.0": { |
| "accuracy": 0.5758, |
| "auroc": 0.5766, |
| "gmean2": 0.3317, |
| "tpr": 0.6042, |
| "tnr": 0.549 |
| }, |
| "50.0": { |
| "accuracy": 0.6061, |
| "auroc": 0.6048, |
| "gmean2": 0.364, |
| "tpr": 0.5625, |
| "tnr": 0.6471 |
| }, |
| "75.0": { |
| "accuracy": 0.5657, |
| "auroc": 0.5656, |
| "gmean2": 0.3199, |
| "tpr": 0.5625, |
| "tnr": 0.5686 |
| }, |
| "100.0": { |
| "accuracy": 0.6162, |
| "auroc": 0.6207, |
| "gmean2": 0.3627, |
| "tpr": 0.7708, |
| "tnr": 0.4706 |
| } |
| } |
| }, |
| "comparisons": { |
| "anthropic--claude-sonnet-4.6_baseline": { |
| "igc": { |
| "10": 0.23760000000000003, |
| "25": 0.27469999999999994, |
| "50": 0.3689, |
| "75": 0.3197, |
| "100": 0.17809999999999993 |
| }, |
| "detection_latency_probe": 10, |
| "detection_latency_baseline": null |
| }, |
| "gemini_baseline": { |
| "igc": { |
| "10": 0.23450000000000004, |
| "25": 0.22139999999999993, |
| "50": 0.25429999999999997, |
| "75": 0.2356999999999999, |
| "100": 0.19709999999999994 |
| }, |
| "detection_latency_probe": 10, |
| "detection_latency_baseline": null |
| }, |
| "google--gemini-3-flash-preview_baseline": { |
| "igc": { |
| "10": 0.23450000000000004, |
| "25": 0.22139999999999993, |
| "50": 0.25429999999999997, |
| "75": 0.2356999999999999, |
| "100": 0.19709999999999994 |
| }, |
| "detection_latency_probe": 10, |
| "detection_latency_baseline": null |
| }, |
| "google--gemini-3.1-pro-preview_baseline": { |
| "igc": { |
| "10": 0.18610000000000004, |
| "25": 0.1473, |
| "50": 0.19179999999999997, |
| "75": 0.20939999999999992, |
| "100": 0.2075999999999999 |
| }, |
| "detection_latency_probe": 10, |
| "detection_latency_baseline": null |
| }, |
| "openai--gpt-5.1_baseline": { |
| "igc": { |
| "10": 0.19099999999999995, |
| "25": 0.21219999999999994, |
| "50": 0.2249, |
| "75": 0.26449999999999996, |
| "100": 0.20939999999999992 |
| }, |
| "detection_latency_probe": 10, |
| "detection_latency_baseline": null |
| } |
| } |
| } |