{"alignment_dimensions_with_observations":["constraint_compliance","robustness"],"alignment_dimensions_without_observations":["authorized_goal_following","non_deception","oversight_and_correction"],"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"content_sha256":"e6809b9c35205eb38e12ae60cb02d941e6a6059c0993d0ce86cc98ebb1a63365","coverage_meaning":"At least one observation, not complete or validated dimension coverage.","evidence_status":"recorded","honesty_dimensions_with_observations":["uncertainty"],"honesty_dimensions_without_observations":["belief_revision","citation_integrity","evidence_support","factual_accuracy","pressure_resistance"],"id":"hm-d4587ee1eeeb8ed0fc0c","integrity_meaning":"Content-addressed record only; not an authenticated issuer signature or independent scientific validation.","lab":"Google","limitations":["Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent."],"mark_status":"not_awarded","model_label":"Gemini 2.5 Flash","observations":[{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Stability of confidence across runs, not correctness of confidence.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"C_conf","metric_label":"Confidence consistency","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":1,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0055936585651638,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.990196985647974},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Repeat agreement on success/failure; always failing can score 1.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"C_out","metric_label":"Outcome consistency","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":8,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.082051282051282,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.7264957264957265},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Stability of resource consumption across repeated runs.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"C_res","metric_label":"Resource consistency","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":15,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0277130071794411,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.7647746523186545},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Similarity of action distributions across successful runs.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"C_traj_d","metric_label":"Action-distribution consistency","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0338416511835162,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.8266372418279712},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Similarity of action order across successful runs.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"C_traj_s","metric_label":"Action-sequence consistency","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":13,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0358719467117686,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.6985537172294635},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"AUROC of confidence distinguishing success from failure.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"P_auroc","metric_label":"Confidence discrimination","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"calibration","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":9,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0307445618269864,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.5258152173913043},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Inverted binary Brier score; higher is better. Not the raw Brier loss.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"P_brier","metric_label":"Brier quality: 1 − Brier","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"calibration","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0541545517704305,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.5999358974358975},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"One minus expected calibration error using source bins.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"P_cal","metric_label":"Calibration: 1 − ECE","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"calibration","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.0548741791021375,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.5987179487179486},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Source normalized excess area under the risk–coverage curve.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"P_rc","metric_label":"Risk–coverage quality","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"calibration","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.0},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Perturbed success / baseline success, capped at 1. Read beside task success.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"R_fault","metric_label":"Tool-failure robustness","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"robustness","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":15,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.1243813287926104,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.7808924485125858},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Rephrased-prompt success / baseline success, capped at 1.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"R_prompt","metric_label":"Prompt robustness","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"robustness","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":14,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.1124794028802897,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.8260869565217391},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Structural-perturbation success / baseline success, capped at 1.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"R_struct","metric_label":"Format robustness","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"robustness","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":13,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":0.1553771178228099,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.8478260869565217},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Fraction of source tasks with zero detected constraint violations.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"S_comp","metric_label":"Constraint compliance","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"constraint_compliance","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":10,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.858974358974359},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"One minus severity conditional on a violation; equals 1 if no violations. Not the probability of avoiding harm.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"S_harm","metric_label":"Conditional severity: inverted","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"constraint_compliance","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":15,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.2727272727272727},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Successful tasks under the source agent setup.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"accuracy","metric_label":"Task success","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"usefulness","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":5.56944604458839,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"percent","value":58.97435897435898},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Source category-weighted consistency aggregate. Consistent failure can score well.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"dim_Consistency","metric_label":"Consistency","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":10,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.7512886194476994},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"HAL uses the inverted Brier score as this aggregate.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"dim_Predictability","metric_label":"Predictability","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"calibration","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.5999358974358975},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Source mean of fault, format and prompt robustness ratios.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"dim_Robustness","metric_label":"Robustness","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"robustness","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":14,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.8182684973302822},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"One minus source violation frequency times conditional severity. Not universal safety.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"dim_Safety","metric_label":"Constraint-risk score","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"constraint_compliance","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.8974358974358975},{"aggregation_blocker":"Agent + scaffold result on the named benchmark. Not a bare-model alignment score. Repeats and overlapping aggregate metrics are dependent.","cohort":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline","confidence_interval":null,"configuration":{"agent_id":"taubench_toolcalling_gemini_2_5_flash","benchmark":"taubench_airline","model_revision":null,"provider_name":"google","reasoning_effort":null,"scaffold":"HAL benchmark-specific agent; source identity retained, not equated with a bare model.","source_agent_slug":"gemini-2-5-flash"},"configuration_id":"hal:a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20:taubench_airline:taubench_toolcalling_gemini_2_5_flash","definition":"Source aggregate of consistency, predictability and robustness. Excludes safety.","denominator":null,"direction":"higher","evaluation_date":null,"evidence_class":"independent_benchmark_reported","lab":"Google","metric":"overall_reliability","metric_label":"HAL reliability","model_label":"Gemini 2.5 Flash","model_release_date":"2025-03-25","numerator":null,"overall_score":null,"pillar":"reliability","profile_id":"hm-d4587ee1eeeb8ed0fc0c","rank":11,"rank_meaning":"Descriptive native-metric order; not a significance or overall honesty rank.","source":"hal","source_runs":3,"source_sha256":"a0da5842bdd0ca4563dd0d9a67efd62e47e24f9ee7fd42090f196e8bd5bc3a20","source_standard_error":null,"source_task_count":26,"source_updated":null,"source_url":"https://hal.cs.princeton.edu/reliability/benchmark/taubench_airline/","split":"source_benchmark_tasks","track":"agent_alignment_evidence","uncertainty_note":"Source standard error retained where supplied; not converted to a confidence interval or a significance rank.","unit":"fraction","value":0.723164338071293}],"overall_alignment_score":null,"overall_honesty_score":null,"priority_gaps":["belief_revision","citation_integrity","evidence_support","factual_accuracy","pressure_resistance"],"requirements":[{"detail":"The standard is a draft; no pass thresholds or award authority have been validated.","id":"standard_ratified","label":"A published, validated standard and scoring policy","status":"not_met"},{"detail":"The current source packet does not establish this requirement.","id":"configuration_pinned","label":"Model revision, provider, scaffold, tools and test settings are pinned","status":"not_established"},{"detail":"Missing or incomplete common-basket coverage. Some observations cover: uncertainty.","id":"complete_honesty_basket","label":"All required honesty dimensions measured on the same declared configuration","status":"not_met"},{"detail":"The current source packet does not establish this requirement.","id":"held_out_evaluation","label":"Held-out evaluation separated from training and development","status":"not_established"},{"detail":"The current source packet does not establish this requirement.","id":"usefulness_and_evasion","label":"Useful answering, abstention and evasion measured together","status":"not_established"},{"detail":"The source reports standard errors for some metrics. No cross-benchmark or award uncertainty is established.","id":"uncertainty_validated","label":"Uncertainty reflects task dependence and supports the claimed comparison","status":"not_established"},{"detail":"The current source packet does not establish this requirement.","id":"independent_review","label":"Independent review reconstructs the evidence and scoring","status":"not_established"},{"detail":"Evidence digests exist. Certification issuance, issuer authentication and renewal are not implemented.","id":"issuer_and_lifecycle","label":"Issuer identity, expiry, re-evaluation and withdrawal process are established","status":"not_met"}],"schema":"honesty-mark-profile/v0.1","scope":"Native source/configuration results only. This profile is not an awarded Honesty Mark.","source":"hal","sources":[{"comparison_caution":"Agent/scaffold results on GAIA and clean airline τ-bench. HAL reliability excludes safety; metric aggregates overlap their components. Source standard errors are not confidence intervals. Model release dates are not evaluation dates.","id":"hal","import_status":"imported","metrics":["Task success","HAL reliability","Consistency","Outcome consistency","Action-distribution consistency","Action-sequence consistency","Resource consistency","Confidence consistency","Predictability","Calibration: 1 − ECE","Confidence discrimination","Brier quality: 1 − Brier","Risk–coverage quality","Robustness","Tool-failure robustness","Format robustness","Prompt robustness","Constraint-risk score","Constraint compliance","Conditional severity: inverted"],"name":"HAL Agent Reliability","organization":"SAgE / Princeton University","pillars":["calibration","constraint_compliance","robustness","reliability"],"reviewed_on":"2026-10-11","url":"https://hal.cs.princeton.edu/reliability/"}],"standard_sha256":"0cec00d16bbb6e2a02a28e43e43394e3eb309523c2ed697913154c4a2e6e6f6c","standard_version":"honesty-mark-standard/v0.1"}