{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Structural resonance of baseplate"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Pump cavitation at low NPSH", "Structural resonance of baseplate", "Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Engineer's handoff: Centrifugal pump at 1694 rpm. Vibration spectrum peak at 56.5 Hz (2x) amplitude 6.6 mm/s, 1x 2.0 mm/s. Axial vibration 6.7 mm/s (radial/axial ratio 0.30). Coupling replaced last week; laser alignment record missing. Bearing temp 73 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1694.0, 6.6, 2.0, 6.7, 73.0, 56.46666666666667, 0.29850746268656714, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Rolling-element bearing outer-race defect", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Rotor mass imbalance", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Pump cavitation at low NPSH"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Structural resonance of baseplate"], "domain": "mechanical", "evidence": "Shift log excerpt (cleansed): Centrifugal pump at 1890 rpm. Vibration spectrum peak at 63.0 Hz (2x) amplitude 8.5 mm/s, 1x 2.3 mm/s. Axial vibration 3.8 mm/s (radial/axial ratio 0.61). Coupling replaced last week; laser alignment record missing. Bearing temp 69 C. Firmware banner printed at boot duplicate log line suppressed x12", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 1, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1890.0, 8.5, 2.3, 3.8, 69.0, 63.0, 0.6052631578947368, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Structural resonance of baseplate", "Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Rotor mass imbalance"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH", "Rotor mass imbalance", "Structural resonance of baseplate"], "domain": "mechanical", "evidence": "Operator note: Centrifugal pump at 2826 rpm. Vibration spectrum peak at 94.2 Hz (2x) amplitude 5.5 mm/s, 1x 2.4 mm/s. Axial vibration 3.4 mm/s (radial/axial ratio 0.71). Coupling replaced last week; laser alignment record missing. Bearing temp 57 C. A colleague suspects 'rotor mass imbalance' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2826.0, 5.5, 2.4, 3.4, 57.0, 94.2, 0.7058823529411765, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Rotor mass imbalance", "Rolling-element bearing outer-race defect"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Rotor mass imbalance", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Pump cavitation at low NPSH"], "domain": "mechanical", "evidence": "Ticket summary: Centrifugal pump at 1533 rpm. Vibration spectrum peak at 51.1 Hz (2x) amplitude 6.0 mm/s, 1x 2.2 mm/s. Axial vibration 4.2 mm/s (radial/axial ratio 0.52). Coupling replaced last week; laser alignment record missing. Bearing temp 71 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1533.0, 6.0, 2.2, 4.2, 71.0, 51.1, 0.5238095238095238, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Rolling-element bearing outer-race defect", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rotor mass imbalance"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Structural resonance of baseplate", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Pump cavitation at low NPSH", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Field report: Centrifugal pump at 2735 rpm. Vibration spectrum peak at 91.2 Hz (2x) amplitude 6.8 mm/s, 1x 1.5 mm/s. Axial vibration 4.2 mm/s (radial/axial ratio 0.36). Coupling replaced last week; laser alignment record missing. Bearing temp 73 C. Firmware banner printed at boot [INFO] heartbeat ok", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2735.0, 6.8, 1.5, 4.2, 73.0, 91.16666666666667, 0.35714285714285715, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Rotor mass imbalance", "Structural resonance of baseplate"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Pump cavitation at low NPSH", "Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Accelerometer mounting looseness", "Structural resonance of baseplate"], "domain": "mechanical", "evidence": "Centrifugal pump at 2923 rpm. Vibration spectrum peak at 97.4 Hz (2x) amplitude 8.9 mm/s, 1x 1.3 mm/s. Axial vibration 4.6 mm/s (radial/axial ratio 0.28). Coupling replaced last week; laser alignment record missing. Bearing temp 65 C. A colleague suspects 'pump cavitation at low npsh' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 3, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2923.0, 8.9, 1.3, 4.6, 65.0, 97.43333333333334, 0.28260869565217395, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Rotor mass imbalance", "Accelerometer mounting looseness", "Pump cavitation at low NPSH"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Pump cavitation at low NPSH", "Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Engineer's handoff: Centrifugal pump at 2620 rpm. Vibration spectrum peak at 87.3 Hz (2x) amplitude 5.8 mm/s, 1x 2.8 mm/s. Axial vibration 4.0 mm/s (radial/axial ratio 0.70). Coupling replaced last week; laser alignment record missing. Bearing temp 64 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2620.0, 5.8, 2.8, 4.0, 64.0, 87.33333333333333, 0.7, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Pump cavitation at low NPSH", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Accelerometer mounting looseness", "Rotor mass imbalance", "Structural resonance of baseplate", "Rolling-element bearing outer-race defect"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH", "Structural resonance of baseplate", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Shift log excerpt (cleansed): Centrifugal pump at 2144 rpm. Vibration spectrum peak at 71.5 Hz (2x) amplitude 6.1 mm/s, 1x 1.9 mm/s. Axial vibration 4.7 mm/s (radial/axial ratio 0.40). Coupling replaced last week; laser alignment record missing. Bearing temp 68 C. Firmware banner printed at boot Note: shift change at 14:00", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 1, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2144.0, 6.1, 1.9, 4.7, 68.0, 71.46666666666667, 0.40425531914893614, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH", "Rotor mass imbalance", "Structural resonance of baseplate", "Shaft misalignment across the new coupling (2x + high axial vibration)"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH", "Structural resonance of baseplate", "Rotor mass imbalance", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Shift log excerpt (cleansed): Centrifugal pump at 1638 rpm. Vibration spectrum peak at 54.6 Hz (2x) amplitude 9.0 mm/s, 1x 1.1 mm/s. Axial vibration 4.2 mm/s (radial/axial ratio 0.26). Coupling replaced last week; laser alignment record missing. Bearing temp 68 C. A colleague suspects 'rotor mass imbalance' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 5, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1638.0, 9.0, 1.1, 4.2, 68.0, 54.6, 0.2619047619047619, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Rotor mass imbalance", "Accelerometer mounting looseness", "Pump cavitation at low NPSH"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Shift log excerpt (cleansed): Centrifugal pump at 2944 rpm. Vibration spectrum peak at 98.1 Hz (2x) amplitude 6.4 mm/s, 1x 2.3 mm/s. Axial vibration 4.7 mm/s (radial/axial ratio 0.49). Coupling replaced last week; laser alignment record missing. Bearing temp 68 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 2, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2944.0, 6.4, 2.3, 4.7, 68.0, 98.13333333333334, 0.4893617021276595, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Rolling-element bearing outer-race defect", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Accelerometer mounting looseness", "Rotor mass imbalance", "Structural resonance of baseplate", "Pump cavitation at low NPSH"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Rolling-element bearing outer-race defect", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rotor mass imbalance"], "domain": "mechanical", "evidence": "Engineer's handoff: Centrifugal pump at 1759 rpm. Vibration spectrum peak at 58.6 Hz (2x) amplitude 4.1 mm/s, 1x 1.4 mm/s. Axial vibration 4.3 mm/s (radial/axial ratio 0.33). Coupling replaced last week; laser alignment record missing. Bearing temp 61 C. badge reader log entry unrelated Firmware banner printed at boot", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 1, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1759.0, 4.1, 1.4, 4.3, 61.0, 58.63333333333333, 0.3255813953488372, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Rolling-element bearing outer-race defect", "Accelerometer mounting looseness", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rotor mass imbalance"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Centrifugal pump at 2420 rpm. Vibration spectrum peak at 80.7 Hz (2x) amplitude 5.8 mm/s, 1x 1.6 mm/s. Axial vibration 4.5 mm/s (radial/axial ratio 0.36). Coupling replaced last week; laser alignment record missing. Bearing temp 60 C. A colleague suspects 'rolling-element bearing outer-race defect' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 2, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2420.0, 5.8, 1.6, 4.5, 60.0, 80.66666666666667, 0.35555555555555557, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Pump cavitation at low NPSH", "Structural resonance of baseplate", "Rotor mass imbalance", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Accelerometer mounting looseness", "Pump cavitation at low NPSH", "Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate"], "domain": "mechanical", "evidence": "Centrifugal pump at 1513 rpm. Vibration spectrum peak at 50.4 Hz (2x) amplitude 7.6 mm/s, 1x 2.2 mm/s. Axial vibration 3.5 mm/s (radial/axial ratio 0.63). Coupling replaced last week; laser alignment record missing. Bearing temp 57 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 3, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1513.0, 7.6, 2.2, 3.5, 57.0, 50.43333333333333, 0.6285714285714287, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Rolling-element bearing outer-race defect", "Rotor mass imbalance"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Rotor mass imbalance", "Pump cavitation at low NPSH"], "domain": "mechanical", "evidence": "Operator note: Centrifugal pump at 2294 rpm. Vibration spectrum peak at 76.5 Hz (2x) amplitude 4.0 mm/s, 1x 1.1 mm/s. Axial vibration 5.4 mm/s (radial/axial ratio 0.20). Coupling replaced last week; laser alignment record missing. Bearing temp 75 C. [INFO] heartbeat ok duplicate log line suppressed x12", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2294.0, 4.0, 1.1, 5.4, 75.0, 76.46666666666667, 0.20370370370370372, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Accelerometer mounting looseness"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Structural resonance of baseplate", "Pump cavitation at low NPSH"], "domain": "mechanical", "evidence": "Engineer's handoff: Centrifugal pump at 2213 rpm. Vibration spectrum peak at 73.8 Hz (2x) amplitude 6.1 mm/s, 1x 2.9 mm/s. Axial vibration 3.5 mm/s (radial/axial ratio 0.83). Coupling replaced last week; laser alignment record missing. Bearing temp 64 C. A colleague suspects 'rolling-element bearing outer-race defect' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 4, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2213.0, 6.1, 2.9, 3.5, 64.0, 73.76666666666667, 0.8285714285714285, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Rotor mass imbalance"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rotor mass imbalance", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect"], "domain": "mechanical", "evidence": "Ticket summary: Centrifugal pump at 2934 rpm. Vibration spectrum peak at 97.8 Hz (2x) amplitude 8.5 mm/s, 1x 2.3 mm/s. Axial vibration 3.2 mm/s (radial/axial ratio 0.72). Coupling replaced last week; laser alignment record missing. Bearing temp 67 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2934.0, 8.5, 2.3, 3.2, 67.0, 97.8, 0.7187499999999999, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Structural resonance of baseplate", "Rotor mass imbalance", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Field report: Centrifugal pump at 1931 rpm. Vibration spectrum peak at 64.4 Hz (2x) amplitude 4.4 mm/s, 1x 2.4 mm/s. Axial vibration 6.6 mm/s (radial/axial ratio 0.36). Coupling replaced last week; laser alignment record missing. Bearing temp 67 C. Firmware banner printed at boot badge reader log entry unrelated", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1931.0, 4.4, 2.4, 6.6, 67.0, 64.36666666666666, 0.36363636363636365, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Structural resonance of baseplate"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Pump cavitation at low NPSH", "Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Field report: Centrifugal pump at 1723 rpm. Vibration spectrum peak at 57.4 Hz (2x) amplitude 5.6 mm/s, 1x 2.4 mm/s. Axial vibration 6.8 mm/s (radial/axial ratio 0.35). Coupling replaced last week; laser alignment record missing. Bearing temp 62 C. A colleague suspects 'rotor mass imbalance' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [1723.0, 5.6, 2.4, 6.8, 62.0, 57.43333333333333, 0.35294117647058826, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Accelerometer mounting looseness", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Rotor mass imbalance", "Pump cavitation at low NPSH"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Rotor mass imbalance", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Pump cavitation at low NPSH"], "domain": "mechanical", "evidence": "Shift log excerpt (cleansed): Centrifugal pump at 2053 rpm. Vibration spectrum peak at 68.4 Hz (2x) amplitude 5.4 mm/s, 1x 2.7 mm/s. Axial vibration 5.5 mm/s (radial/axial ratio 0.49). Coupling replaced last week; laser alignment record missing. Bearing temp 58 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 1, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2053.0, 5.4, 2.7, 5.5, 58.0, 68.43333333333334, 0.49090909090909096, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Structural resonance of baseplate", "Rotor mass imbalance", "Shaft misalignment across the new coupling (2x + high axial vibration)", "Rolling-element bearing outer-race defect"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Rotor mass imbalance", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Structural resonance of baseplate", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Field report: Centrifugal pump at 2349 rpm. Vibration spectrum peak at 78.3 Hz (2x) amplitude 7.9 mm/s, 1x 1.1 mm/s. Axial vibration 5.8 mm/s (radial/axial ratio 0.19). Coupling replaced last week; laser alignment record missing. Bearing temp 58 C. [INFO] heartbeat ok Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 4, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2349.0, 7.9, 1.1, 5.8, 58.0, 78.3, 0.18965517241379312, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Accelerometer mounting looseness", "Rotor mass imbalance", "Rolling-element bearing outer-race defect", "Pump cavitation at low NPSH", "Structural resonance of baseplate", "Shaft misalignment across the new coupling (2x + high axial vibration)"], "difficulty": "hard", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Rotor mass imbalance", "Structural resonance of baseplate", "Accelerometer mounting looseness"], "domain": "mechanical", "evidence": "Field report: Centrifugal pump at 2771 rpm. Vibration spectrum peak at 92.4 Hz (2x) amplitude 6.7 mm/s, 1x 2.5 mm/s. Axial vibration 6.7 mm/s (radial/axial ratio 0.37). Coupling replaced last week; laser alignment record missing. Bearing temp 57 C. A colleague suspects 'rotor mass imbalance' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 5, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2771.0, 6.7, 2.5, 6.7, 57.0, 92.36666666666666, 0.3731343283582089, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Shaft misalignment across the new coupling (2x + high axial vibration)", "Structural resonance of baseplate", "Pump cavitation at low NPSH", "Accelerometer mounting looseness", "Rolling-element bearing outer-race defect", "Rotor mass imbalance"], "difficulty": "medium", "discriminating_followups": ["Is there an alignment record after the coupling swap?", "Is the 2x peak exactly synchronous or slightly off (BPFO)?", "Does phase across the coupling differ by ~180 deg axially?"], "distractors": ["Structural resonance of baseplate", "Pump cavitation at low NPSH", "Rolling-element bearing outer-race defect", "Accelerometer mounting looseness", "Rotor mass imbalance"], "domain": "mechanical", "evidence": "Field report: Centrifugal pump at 2639 rpm. Vibration spectrum peak at 88.0 Hz (2x) amplitude 4.9 mm/s, 1x 2.3 mm/s. Axial vibration 4.9 mm/s (radial/axial ratio 0.47). Coupling replaced last week; laser alignment record missing. Bearing temp 63 C.", "failure_category": "reagent_material_flaw", "gold_cause": "Shaft misalignment across the new coupling (2x + high axial vibration)", "gold_hypothesis_index": 0, "group_id": "fam:MECH-BEARING-MISALIGN", "id": "SYN-MECH-BEARING-MISALIGN-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Misalignment shows 2x running speed with high axial component; imbalance shows mostly 1x radial.", "Bearing defects show non-synchronous frequencies (BPFO/BPFI), not exact 2x."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "rotating_machinery", "tabular_stats": [2639.0, 4.9, 2.3, 4.9, 63.0, 87.96666666666667, 0.4693877551020407, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump vibration at 2x running speed", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["ADC aliasing of a high-frequency signal", "Firmware sample-rate bug", "Supply regulator failure in the sensor", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Process actually oscillating", "Sensor physically vibrating at mains frequency"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Supply regulator failure in the sensor", "ADC aliasing of a high-frequency signal", "Process actually oscillating", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency"], "domain": "electrical", "evidence": "Shift log excerpt (cleansed): Analog channel shows 59 Hz ripple of 54 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 40 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 3, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [59.0, 54.0, 40.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Sensor physically vibrating at mains frequency", "Firmware sample-rate bug", "Supply regulator failure in the sensor", "Process actually oscillating", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "ADC aliasing of a high-frequency signal"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Supply regulator failure in the sensor", "Process actually oscillating", "Sensor physically vibrating at mains frequency", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Analog channel shows 57 Hz ripple of 47 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 10 m. HMI screensaver restarted Firmware banner printed at boot", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 4, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [57.0, 47.0, 10.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Ground loop through double-ended shield grounding coupling mains current into the measurement", "ADC aliasing of a high-frequency signal", "Process actually oscillating", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "ADC aliasing of a high-frequency signal", "Process actually oscillating"], "domain": "electrical", "evidence": "Shift log excerpt (cleansed): Analog channel shows 54 Hz ripple of 15 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 25 m. A colleague suspects 'sensor physically vibrating at mains frequency' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 0, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [54.0, 15.0, 25.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["ADC aliasing of a high-frequency signal", "Process actually oscillating", "Supply regulator failure in the sensor", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "ADC aliasing of a high-frequency signal", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "Process actually oscillating"], "domain": "electrical", "evidence": "Analog channel shows 52 Hz ripple of 30 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 38 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 3, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [52.0, 30.0, 38.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["ADC aliasing of a high-frequency signal", "Sensor physically vibrating at mains frequency", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Supply regulator failure in the sensor", "Process actually oscillating", "Firmware sample-rate bug"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Supply regulator failure in the sensor", "Process actually oscillating", "ADC aliasing of a high-frequency signal", "Sensor physically vibrating at mains frequency"], "domain": "electrical", "evidence": "Ticket summary: Analog channel shows 60 Hz ripple of 72 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 7 m. HMI screensaver restarted Note: shift change at 14:00", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 2, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [60.0, 72.0, 7.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Sensor physically vibrating at mains frequency", "Process actually oscillating", "Firmware sample-rate bug", "Supply regulator failure in the sensor", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "ADC aliasing of a high-frequency signal"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "ADC aliasing of a high-frequency signal", "Firmware sample-rate bug"], "domain": "electrical", "evidence": "Operator note: Analog channel shows 55 Hz ripple of 33 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 37 m. A colleague suspects 'adc aliasing of a high-frequency signal' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 4, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [55.0, 33.0, 37.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Ground loop through double-ended shield grounding coupling mains current into the measurement", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "ADC aliasing of a high-frequency signal", "Process actually oscillating", "Firmware sample-rate bug"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Operator note: Analog channel shows 59 Hz ripple of 33 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 34 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 0, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [59.0, 33.0, 34.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process actually oscillating", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "ADC aliasing of a high-frequency signal", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Firmware sample-rate bug"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Field report: Analog channel shows 58 Hz ripple of 70 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 20 m. HMI screensaver restarted Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 4, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [58.0, 70.0, 20.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Firmware sample-rate bug", "ADC aliasing of a high-frequency signal", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "Supply regulator failure in the sensor"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Ticket summary: Analog channel shows 58 Hz ripple of 45 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 40 m. A colleague suspects 'supply regulator failure in the sensor' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 2, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [58.0, 45.0, 40.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process actually oscillating", "ADC aliasing of a high-frequency signal", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Firmware sample-rate bug"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Engineer's handoff: Analog channel shows 59 Hz ripple of 20 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 25 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 2, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [59.0, 20.0, 25.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Ground loop through double-ended shield grounding coupling mains current into the measurement", "Sensor physically vibrating at mains frequency", "Firmware sample-rate bug", "Supply regulator failure in the sensor", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "Process actually oscillating"], "domain": "electrical", "evidence": "Ticket summary: Analog channel shows 50 Hz ripple of 47 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 15 m. [INFO] heartbeat ok Firmware banner printed at boot", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 0, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [50.0, 47.0, 15.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Supply regulator failure in the sensor", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "ADC aliasing of a high-frequency signal", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "Process actually oscillating"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Process actually oscillating", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Analog channel shows 56 Hz ripple of 50 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 39 m. A colleague suspects 'sensor physically vibrating at mains frequency' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 1, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [56.0, 50.0, 39.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Process actually oscillating", "ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "Process actually oscillating", "Firmware sample-rate bug", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Engineer's handoff: Analog channel shows 52 Hz ripple of 28 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 34 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 2, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [52.0, 28.0, 34.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Ground loop through double-ended shield grounding coupling mains current into the measurement", "ADC aliasing of a high-frequency signal", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Process actually oscillating"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Operator note: Analog channel shows 52 Hz ripple of 15 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 8 m. [INFO] heartbeat ok Note: shift change at 14:00", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 0, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [52.0, 15.0, 8.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process actually oscillating", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Firmware sample-rate bug", "ADC aliasing of a high-frequency signal", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Field report: Analog channel shows 53 Hz ripple of 75 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 32 m. A colleague suspects 'sensor physically vibrating at mains frequency' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 1, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [53.0, 75.0, 32.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Supply regulator failure in the sensor", "Process actually oscillating", "ADC aliasing of a high-frequency signal", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Supply regulator failure in the sensor", "Process actually oscillating", "ADC aliasing of a high-frequency signal", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency"], "domain": "electrical", "evidence": "Operator note: Analog channel shows 51 Hz ripple of 16 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 15 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 3, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [51.0, 16.0, 15.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Supply regulator failure in the sensor", "Process actually oscillating", "Sensor physically vibrating at mains frequency", "Firmware sample-rate bug", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "ADC aliasing of a high-frequency signal"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Supply regulator failure in the sensor", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "Process actually oscillating", "ADC aliasing of a high-frequency signal"], "domain": "electrical", "evidence": "Ticket summary: Analog channel shows 56 Hz ripple of 53 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 33 m. Note: shift change at 14:00 badge reader log entry unrelated", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 4, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [56.0, 53.0, 33.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Process actually oscillating", "Sensor physically vibrating at mains frequency"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Process actually oscillating", "ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor", "Sensor physically vibrating at mains frequency", "Firmware sample-rate bug"], "domain": "electrical", "evidence": "Operator note: Analog channel shows 58 Hz ripple of 68 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 31 m. A colleague suspects 'adc aliasing of a high-frequency signal' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 3, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [58.0, 68.0, 31.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["ADC aliasing of a high-frequency signal", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Process actually oscillating", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Firmware sample-rate bug", "Process actually oscillating", "Sensor physically vibrating at mains frequency", "ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor"], "domain": "electrical", "evidence": "Operator note: Analog channel shows 52 Hz ripple of 59 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 38 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 1, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [52.0, 59.0, 38.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Supply regulator failure in the sensor", "Process actually oscillating", "Firmware sample-rate bug", "Sensor physically vibrating at mains frequency", "ADC aliasing of a high-frequency signal", "Ground loop through double-ended shield grounding coupling mains current into the measurement"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Process actually oscillating", "Sensor physically vibrating at mains frequency", "Firmware sample-rate bug", "ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor"], "domain": "electrical", "evidence": "Shift log excerpt (cleansed): Analog channel shows 51 Hz ripple of 33 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 38 m. [INFO] heartbeat ok HMI screensaver restarted", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 5, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [51.0, 33.0, 38.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Sensor physically vibrating at mains frequency", "Firmware sample-rate bug", "Process actually oscillating", "ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor", "Ground loop through double-ended shield grounding coupling mains current into the measurement"], "difficulty": "medium", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["Sensor physically vibrating at mains frequency", "ADC aliasing of a high-frequency signal", "Firmware sample-rate bug", "Process actually oscillating", "Supply regulator failure in the sensor"], "domain": "electrical", "evidence": "Analog channel shows 54 Hz ripple of 22 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 16 m. A colleague suspects 'adc aliasing of a high-frequency signal' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 5, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [54.0, 22.0, 16.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process actually oscillating", "ADC aliasing of a high-frequency signal", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Ground loop through double-ended shield grounding coupling mains current into the measurement", "Sensor physically vibrating at mains frequency"], "difficulty": "easy", "discriminating_followups": ["Does lifting one shield end remove the ripple?", "Is ripple exactly at mains frequency and its harmonics?", "Does a differential/isolated input remove it?"], "distractors": ["ADC aliasing of a high-frequency signal", "Sensor physically vibrating at mains frequency", "Supply regulator failure in the sensor", "Firmware sample-rate bug", "Process actually oscillating"], "domain": "electrical", "evidence": "Analog channel shows 53 Hz ripple of 76 mV plus harmonics. Ripple disappears when laptop runs on battery. Shield grounded at both ends; cable length 22 m.", "failure_category": "instrumentation_artifact", "gold_cause": "Ground loop through double-ended shield grounding coupling mains current into the measurement", "gold_hypothesis_index": 4, "group_id": "fam:ELEC-GROUND-LOOP", "id": "SYN-ELEC-GROUND-LOOP-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Disconnecting a ground path (battery power) breaks the loop current.", "Aliasing would not depend on laptop power source."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "signal_integrity", "tabular_stats": [53.0, 76.0, 22.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Mains-frequency noise on sensor channel", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Gripper never closed (width stuck)", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Premature gripper open at lift (control logic)"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout"], "domain": "robotics", "evidence": "Field report: Parallel gripper closes to width 0.039 m (object 0.039 m). Grip force 18 N; object mass 1.09 kg; lift acceleration 2.7 m/s2. Required normal force for no-slip at mu=0.17: 40.1 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 4, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.039, 18.0, 1.09, 2.7, 0.17, 0.039, 40.10558823529412, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Camera misdetection of object pose", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Premature gripper open at lift (control logic)", "Camera misdetection of object pose", "Trajectory too slow causing timeout", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection"], "domain": "robotics", "evidence": "Parallel gripper closes to width 0.027 m (object 0.027 m). Grip force 16 N; object mass 1.25 kg; lift acceleration 4.0 m/s2. Required normal force for no-slip at mu=0.2: 43.2 N. Gripper pads replaced with smoother pads last week. [INFO] heartbeat ok Note: shift change at 14:00", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 1, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.027, 16.0, 1.25, 4.0, 0.2, 0.027, 43.15624999999999, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Camera misdetection of object pose", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection"], "domain": "robotics", "evidence": "Engineer's handoff: Parallel gripper closes to width 0.044 m (object 0.044 m). Grip force 17 N; object mass 0.79 kg; lift acceleration 3.2 m/s2. Required normal force for no-slip at mu=0.29: 17.7 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'premature gripper open at lift (control logic)' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 1, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.044, 17.0, 0.79, 3.2, 0.29, 0.044, 17.720517241379316, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Premature gripper open at lift (control logic)", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Gripper never closed (width stuck)", "Trajectory too slow causing timeout"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Camera misdetection of object pose", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout"], "domain": "robotics", "evidence": "Ticket summary: Parallel gripper closes to width 0.037 m (object 0.037 m). Grip force 13 N; object mass 0.7 kg; lift acceleration 2.0 m/s2. Required normal force for no-slip at mu=0.2: 20.7 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 3, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.037, 13.0, 0.7, 2.0, 0.2, 0.037, 20.667499999999997, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Camera misdetection of object pose", "Gripper never closed (width stuck)", "Trajectory too slow causing timeout", "Wrist F/T sensor offset causing false slip detection", "Premature gripper open at lift (control logic)"], "domain": "robotics", "evidence": "Shift log excerpt (cleansed): Parallel gripper closes to width 0.034 m (object 0.034 m). Grip force 18 N; object mass 0.93 kg; lift acceleration 1.7 m/s2. Required normal force for no-slip at mu=0.18: 29.7 N. Gripper pads replaced with smoother pads last week. Firmware banner printed at boot HMI screensaver restarted", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 5, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.034, 18.0, 0.93, 1.7, 0.18, 0.034, 29.734166666666667, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)"], "domain": "robotics", "evidence": "Field report: Parallel gripper closes to width 0.042 m (object 0.042 m). Grip force 15 N; object mass 1.02 kg; lift acceleration 1.6 m/s2. Required normal force for no-slip at mu=0.26: 22.4 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'wrist f/t sensor offset causing false slip detection' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 4, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.042, 15.0, 1.02, 1.6, 0.26, 0.042, 22.381153846153847, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)"], "domain": "robotics", "evidence": "Engineer's handoff: Parallel gripper closes to width 0.039 m (object 0.039 m). Grip force 13 N; object mass 0.76 kg; lift acceleration 2.8 m/s2. Required normal force for no-slip at mu=0.25: 19.2 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 2, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.039, 13.0, 0.76, 2.8, 0.25, 0.039, 19.1672, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Camera misdetection of object pose"], "domain": "robotics", "evidence": "Parallel gripper closes to width 0.031 m (object 0.031 m). Grip force 15 N; object mass 0.83 kg; lift acceleration 1.5 m/s2. Required normal force for no-slip at mu=0.21: 22.4 N. Gripper pads replaced with smoother pads last week. badge reader log entry unrelated duplicate log line suppressed x12", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 0, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.031, 15.0, 0.83, 1.5, 0.21, 0.031, 22.350714285714286, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Wrist F/T sensor offset causing false slip detection", "Camera misdetection of object pose", "Gripper never closed (width stuck)", "Trajectory too slow causing timeout", "Premature gripper open at lift (control logic)"], "domain": "robotics", "evidence": "Shift log excerpt (cleansed): Parallel gripper closes to width 0.03 m (object 0.03 m). Grip force 14 N; object mass 0.9 kg; lift acceleration 3.8 m/s2. Required normal force for no-slip at mu=0.29: 21.1 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'gripper never closed (width stuck)' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 2, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.03, 14.0, 0.9, 3.8, 0.29, 0.03, 21.11896551724138, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Gripper never closed (width stuck)", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection"], "domain": "robotics", "evidence": "Operator note: Parallel gripper closes to width 0.021 m (object 0.021 m). Grip force 17 N; object mass 1.34 kg; lift acceleration 2.6 m/s2. Required normal force for no-slip at mu=0.23: 36.2 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 3, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.021, 17.0, 1.34, 2.6, 0.23, 0.021, 36.15086956521739, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Camera misdetection of object pose", "Premature gripper open at lift (control logic)"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Trajectory too slow causing timeout", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Premature gripper open at lift (control logic)", "Camera misdetection of object pose"], "domain": "robotics", "evidence": "Parallel gripper closes to width 0.029 m (object 0.029 m). Grip force 15 N; object mass 1.21 kg; lift acceleration 1.9 m/s2. Required normal force for no-slip at mu=0.22: 32.2 N. Gripper pads replaced with smoother pads last week. Note: shift change at 14:00 [INFO] heartbeat ok", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 0, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.029, 15.0, 1.21, 1.9, 0.22, 0.029, 32.2025, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Wrist F/T sensor offset causing false slip detection", "Camera misdetection of object pose"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Camera misdetection of object pose", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Wrist F/T sensor offset causing false slip detection"], "domain": "robotics", "evidence": "Engineer's handoff: Parallel gripper closes to width 0.035 m (object 0.035 m). Grip force 15 N; object mass 0.89 kg; lift acceleration 2.8 m/s2. Required normal force for no-slip at mu=0.15: 37.4 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'wrist f/t sensor offset causing false slip detection' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 0, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.035, 15.0, 0.89, 2.8, 0.15, 0.035, 37.409666666666666, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Trajectory too slow causing timeout", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Camera misdetection of object pose", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Camera misdetection of object pose", "Trajectory too slow causing timeout", "Premature gripper open at lift (control logic)"], "domain": "robotics", "evidence": "Operator note: Parallel gripper closes to width 0.047 m (object 0.047 m). Grip force 13 N; object mass 0.54 kg; lift acceleration 2.1 m/s2. Required normal force for no-slip at mu=0.19: 16.9 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 2, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.047, 13.0, 0.54, 2.1, 0.19, 0.047, 16.924736842105265, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Camera misdetection of object pose", "Trajectory too slow causing timeout"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Premature gripper open at lift (control logic)", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Camera misdetection of object pose", "Trajectory too slow causing timeout"], "domain": "robotics", "evidence": "Operator note: Parallel gripper closes to width 0.029 m (object 0.029 m). Grip force 14 N; object mass 1.08 kg; lift acceleration 1.2 m/s2. Required normal force for no-slip at mu=0.3: 19.8 N. Gripper pads replaced with smoother pads last week. badge reader log entry unrelated Note: shift change at 14:00", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 3, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.029, 14.0, 1.08, 1.2, 0.3, 0.029, 19.818, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Gripper never closed (width stuck)", "Premature gripper open at lift (control logic)", "Camera misdetection of object pose"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Trajectory too slow causing timeout", "Premature gripper open at lift (control logic)", "Camera misdetection of object pose", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection"], "domain": "robotics", "evidence": "Ticket summary: Parallel gripper closes to width 0.021 m (object 0.021 m). Grip force 15 N; object mass 1.11 kg; lift acceleration 2.7 m/s2. Required normal force for no-slip at mu=0.28: 24.8 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'premature gripper open at lift (control logic)' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 2, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.021, 15.0, 1.11, 2.7, 0.28, 0.021, 24.796607142857145, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Gripper never closed (width stuck)", "Camera misdetection of object pose", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Trajectory too slow causing timeout", "Wrist F/T sensor offset causing false slip detection", "Premature gripper open at lift (control logic)"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Premature gripper open at lift (control logic)", "Camera misdetection of object pose", "Trajectory too slow causing timeout"], "domain": "robotics", "evidence": "Ticket summary: Parallel gripper closes to width 0.047 m (object 0.047 m). Grip force 19 N; object mass 1.46 kg; lift acceleration 2.9 m/s2. Required normal force for no-slip at mu=0.22: 42.2 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 2, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.047, 19.0, 1.46, 2.9, 0.22, 0.047, 42.17409090909091, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Premature gripper open at lift (control logic)", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Trajectory too slow causing timeout", "Camera misdetection of object pose"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Premature gripper open at lift (control logic)"], "domain": "robotics", "evidence": "Shift log excerpt (cleansed): Parallel gripper closes to width 0.042 m (object 0.042 m). Grip force 10 N; object mass 1.34 kg; lift acceleration 3.9 m/s2. Required normal force for no-slip at mu=0.24: 38.3 N. Gripper pads replaced with smoother pads last week. Unrelated: cafeteria chiller alarm acknowledged HMI screensaver restarted", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 0, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.042, 10.0, 1.34, 3.9, 0.24, 0.042, 38.27375000000001, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Gripper never closed (width stuck)", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Premature gripper open at lift (control logic)"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Gripper never closed (width stuck)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Premature gripper open at lift (control logic)"], "domain": "robotics", "evidence": "Ticket summary: Parallel gripper closes to width 0.036 m (object 0.036 m). Grip force 16 N; object mass 0.9 kg; lift acceleration 2.5 m/s2. Required normal force for no-slip at mu=0.25: 22.2 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'premature gripper open at lift (control logic)' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 4, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.036, 16.0, 0.9, 2.5, 0.25, 0.036, 22.158, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Trajectory too slow causing timeout", "Gripper never closed (width stuck)", "Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Wrist F/T sensor offset causing false slip detection"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout"], "domain": "robotics", "evidence": "Engineer's handoff: Parallel gripper closes to width 0.038 m (object 0.038 m). Grip force 11 N; object mass 0.84 kg; lift acceleration 3.7 m/s2. Required normal force for no-slip at mu=0.28: 20.3 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 4, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.038, 11.0, 0.84, 3.7, 0.28, 0.038, 20.265, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Wrist F/T sensor offset causing false slip detection", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Gripper never closed (width stuck)", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Camera misdetection of object pose", "Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)"], "domain": "robotics", "evidence": "Field report: Parallel gripper closes to width 0.048 m (object 0.048 m). Grip force 18 N; object mass 1.11 kg; lift acceleration 3.6 m/s2. Required normal force for no-slip at mu=0.3: 24.8 N. Gripper pads replaced with smoother pads last week. Unrelated: cafeteria chiller alarm acknowledged duplicate log line suppressed x12", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 5, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.048, 18.0, 1.11, 3.6, 0.3, 0.048, 24.808500000000002, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Premature gripper open at lift (control logic)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Gripper never closed (width stuck)", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)"], "difficulty": "hard", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)", "Trajectory too slow causing timeout", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection"], "domain": "robotics", "evidence": "Shift log excerpt (cleansed): Parallel gripper closes to width 0.044 m (object 0.044 m). Grip force 18 N; object mass 0.58 kg; lift acceleration 3.3 m/s2. Required normal force for no-slip at mu=0.21: 18.1 N. Gripper pads replaced with smoother pads last week. A colleague suspects 'wrist f/t sensor offset causing false slip detection' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 5, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.044, 18.0, 0.58, 3.3, 0.21, 0.044, 18.104285714285712, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Gripper never closed (width stuck)", "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "Premature gripper open at lift (control logic)", "Camera misdetection of object pose", "Wrist F/T sensor offset causing false slip detection", "Trajectory too slow causing timeout"], "difficulty": "medium", "discriminating_followups": ["Does commanded width stay closed while object slides?", "What is pad friction coefficient vs old pads?", "Does halving lift acceleration fix it?"], "distractors": ["Wrist F/T sensor offset causing false slip detection", "Camera misdetection of object pose", "Trajectory too slow causing timeout", "Premature gripper open at lift (control logic)", "Gripper never closed (width stuck)"], "domain": "robotics", "evidence": "Parallel gripper closes to width 0.045 m (object 0.045 m). Grip force 16 N; object mass 0.58 kg; lift acceleration 1.4 m/s2. Required normal force for no-slip at mu=0.26: 12.5 N. Gripper pads replaced with smoother pads last week.", "failure_category": "reagent_material_flaw", "gold_cause": "Insufficient friction from new pads: grip force below no-slip requirement m(g+a)/(2mu)", "gold_hypothesis_index": 1, "group_id": "fam:ROB-GRASP-FRICTION", "id": "SYN-ROB-GRASP-FRICTION-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["No-slip requires 2*mu*Fn >= m*(g+a).", "Premature open shows commanded width jumping to max at lift; slip shows gradual width/force change at constant command."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "manipulation", "tabular_stats": [0.045, 16.0, 0.58, 1.4, 0.26, 0.045, 12.503461538461538, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Object slips during lift", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Residual stress relief", "Hardness tester miscalibration", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Porosity from casting", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Residual stress relief", "Grain coarsening (Hall-Petch)", "Hardness tester miscalibration", "Porosity from casting", "Hydrogen contamination"], "domain": "materials", "evidence": "Shift log excerpt (cleansed): Alloy yield strength dropped 11% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (54 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 5, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [11.0, 54.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Grain coarsening (Hall-Petch)", "Porosity from casting", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hardness tester miscalibration", "Hydrogen contamination", "Residual stress relief"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Hydrogen contamination", "Porosity from casting", "Residual stress relief", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)"], "domain": "materials", "evidence": "Alloy yield strength dropped 16% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (36 um before and after). TEM shows coarsened precipitates. [INFO] heartbeat ok Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 2, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [16.0, 36.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hardness tester miscalibration", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Porosity from casting", "Residual stress relief"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Grain coarsening (Hall-Petch)", "Residual stress relief", "Porosity from casting", "Hardness tester miscalibration", "Hydrogen contamination"], "domain": "materials", "evidence": "Alloy yield strength dropped 22% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (39 um before and after). TEM shows coarsened precipitates. A colleague suspects 'hydrogen contamination' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 1, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [22.0, 39.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Residual stress relief", "Porosity from casting", "Hardness tester miscalibration"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Hardness tester miscalibration", "Residual stress relief", "Porosity from casting", "Grain coarsening (Hall-Petch)", "Hydrogen contamination"], "domain": "materials", "evidence": "Alloy yield strength dropped 15% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (60 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 2, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [15.0, 60.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Porosity from casting", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hardness tester miscalibration", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Residual stress relief"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Porosity from casting", "Hydrogen contamination", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Residual stress relief"], "domain": "materials", "evidence": "Operator note: Alloy yield strength dropped 12% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (42 um before and after). TEM shows coarsened precipitates. HMI screensaver restarted Firmware banner printed at boot", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 1, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [12.0, 42.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hydrogen contamination", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Residual stress relief", "Porosity from casting"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Hydrogen contamination", "Porosity from casting", "Residual stress relief", "Grain coarsening (Hall-Petch)", "Hardness tester miscalibration"], "domain": "materials", "evidence": "Alloy yield strength dropped 12% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (29 um before and after). TEM shows coarsened precipitates. A colleague suspects 'grain coarsening (hall-petch)' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 3, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [12.0, 29.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hydrogen contamination", "Porosity from casting", "Hardness tester miscalibration", "Residual stress relief", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Grain coarsening (Hall-Petch)"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Residual stress relief", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Hydrogen contamination", "Porosity from casting"], "domain": "materials", "evidence": "Operator note: Alloy yield strength dropped 14% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (51 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 4, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [14.0, 51.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Porosity from casting", "Hardness tester miscalibration", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Residual stress relief"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Porosity from casting", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Residual stress relief", "Hydrogen contamination"], "domain": "materials", "evidence": "Alloy yield strength dropped 22% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (46 um before and after). TEM shows coarsened precipitates. HMI screensaver restarted duplicate log line suppressed x12", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 0, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [22.0, 46.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Residual stress relief", "Grain coarsening (Hall-Petch)", "Hardness tester miscalibration", "Porosity from casting", "Hydrogen contamination", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Residual stress relief", "Hardness tester miscalibration", "Porosity from casting", "Hydrogen contamination", "Grain coarsening (Hall-Petch)"], "domain": "materials", "evidence": "Alloy yield strength dropped 22% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (37 um before and after). TEM shows coarsened precipitates. A colleague suspects 'hydrogen contamination' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 5, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [22.0, 37.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hardness tester miscalibration", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Grain coarsening (Hall-Petch)", "Residual stress relief", "Porosity from casting", "Hydrogen contamination"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Grain coarsening (Hall-Petch)", "Residual stress relief", "Porosity from casting", "Hardness tester miscalibration", "Hydrogen contamination"], "domain": "materials", "evidence": "Operator note: Alloy yield strength dropped 17% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (27 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 1, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [17.0, 27.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Residual stress relief", "Porosity from casting", "Hardness tester miscalibration"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Hydrogen contamination", "Residual stress relief", "Porosity from casting"], "domain": "materials", "evidence": "Field report: Alloy yield strength dropped 23% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (55 um before and after). TEM shows coarsened precipitates. duplicate log line suppressed x12 HMI screensaver restarted", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 0, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [23.0, 55.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Residual stress relief", "Hydrogen contamination", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Grain coarsening (Hall-Petch)", "Porosity from casting", "Hardness tester miscalibration"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Grain coarsening (Hall-Petch)", "Hydrogen contamination", "Hardness tester miscalibration", "Residual stress relief", "Porosity from casting"], "domain": "materials", "evidence": "Alloy yield strength dropped 11% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (54 um before and after). TEM shows coarsened precipitates. A colleague suspects 'hydrogen contamination' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 2, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [11.0, 54.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Porosity from casting", "Hydrogen contamination", "Residual stress relief"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Grain coarsening (Hall-Petch)", "Porosity from casting", "Hydrogen contamination", "Hardness tester miscalibration", "Residual stress relief"], "domain": "materials", "evidence": "Shift log excerpt (cleansed): Alloy yield strength dropped 18% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (27 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 0, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [18.0, 27.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Porosity from casting", "Residual stress relief", "Hardness tester miscalibration", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Residual stress relief", "Hardness tester miscalibration", "Porosity from casting", "Grain coarsening (Hall-Petch)", "Hydrogen contamination"], "domain": "materials", "evidence": "Operator note: Alloy yield strength dropped 20% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (37 um before and after). TEM shows coarsened precipitates. duplicate log line suppressed x12 Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 5, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [20.0, 37.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hydrogen contamination", "Hardness tester miscalibration", "Residual stress relief", "Porosity from casting", "Grain coarsening (Hall-Petch)"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Grain coarsening (Hall-Petch)", "Porosity from casting", "Hydrogen contamination", "Hardness tester miscalibration", "Residual stress relief"], "domain": "materials", "evidence": "Engineer's handoff: Alloy yield strength dropped 17% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (36 um before and after). TEM shows coarsened precipitates. A colleague suspects 'hardness tester miscalibration' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 0, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [17.0, 36.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Residual stress relief", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Hardness tester miscalibration", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Porosity from casting"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Hydrogen contamination", "Porosity from casting", "Grain coarsening (Hall-Petch)", "Residual stress relief", "Hardness tester miscalibration"], "domain": "materials", "evidence": "Operator note: Alloy yield strength dropped 20% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (44 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 4, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [20.0, 44.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Porosity from casting", "Hardness tester miscalibration", "Residual stress relief", "Grain coarsening (Hall-Petch)", "Hydrogen contamination"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Grain coarsening (Hall-Petch)", "Hardness tester miscalibration", "Porosity from casting", "Hydrogen contamination", "Residual stress relief"], "domain": "materials", "evidence": "Field report: Alloy yield strength dropped 24% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (54 um before and after). TEM shows coarsened precipitates. Unrelated: cafeteria chiller alarm acknowledged [INFO] heartbeat ok", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 0, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [24.0, 54.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Grain coarsening (Hall-Petch)", "Hardness tester miscalibration", "Hydrogen contamination", "Residual stress relief", "Porosity from casting", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Residual stress relief", "Hardness tester miscalibration", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Porosity from casting"], "domain": "materials", "evidence": "Engineer's handoff: Alloy yield strength dropped 20% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (32 um before and after). TEM shows coarsened precipitates. A colleague suspects 'hydrogen contamination' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 5, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [20.0, 32.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Grain coarsening (Hall-Petch)", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Hardness tester miscalibration", "Residual stress relief", "Porosity from casting", "Hydrogen contamination"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Porosity from casting", "Hardness tester miscalibration", "Residual stress relief", "Grain coarsening (Hall-Petch)", "Hydrogen contamination"], "domain": "materials", "evidence": "Operator note: Alloy yield strength dropped 19% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (44 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 1, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [19.0, 44.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Porosity from casting", "Hardness tester miscalibration", "Hydrogen contamination", "Grain coarsening (Hall-Petch)", "Residual stress relief"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Porosity from casting", "Hardness tester miscalibration", "Grain coarsening (Hall-Petch)", "Residual stress relief", "Hydrogen contamination"], "domain": "materials", "evidence": "Engineer's handoff: Alloy yield strength dropped 11% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (23 um before and after). TEM shows coarsened precipitates. Firmware banner printed at boot Note: shift change at 14:00", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 0, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [11.0, 23.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Hardness tester miscalibration", "Hydrogen contamination", "Porosity from casting", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Grain coarsening (Hall-Petch)", "Residual stress relief"], "difficulty": "hard", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Porosity from casting", "Grain coarsening (Hall-Petch)", "Hardness tester miscalibration", "Residual stress relief", "Hydrogen contamination"], "domain": "materials", "evidence": "Engineer's handoff: Alloy yield strength dropped 15% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (50 um before and after). TEM shows coarsened precipitates. A colleague suspects 'hardness tester miscalibration' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 3, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [15.0, 50.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Residual stress relief", "Hydrogen contamination", "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "Grain coarsening (Hall-Petch)", "Porosity from casting", "Hardness tester miscalibration"], "difficulty": "medium", "discriminating_followups": ["Hardness vs aging time curve - peak then fall?", "Precipitate size distribution before/after?", "Does re-solution + peak age restore strength?"], "distractors": ["Hardness tester miscalibration", "Residual stress relief", "Hydrogen contamination", "Porosity from casting", "Grain coarsening (Hall-Petch)"], "domain": "materials", "evidence": "Alloy yield strength dropped 16% after a longer age. Team blames grain growth (Hall-Petch). Measured grain size unchanged (37 um before and after). TEM shows coarsened precipitates.", "failure_category": "flawed_hypothesis", "gold_cause": "Overaging: precipitate coarsening reduced precipitation strengthening, not grain growth", "gold_hypothesis_index": 2, "group_id": "fam:MAT-HALL-PETCH", "id": "SYN-MAT-HALL-PETCH-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Hall-Petch depends on grain size, which did not change.", "Orowan strengthening falls as precipitates coarsen and spacing grows."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "metallurgy", "tabular_stats": [16.0, 37.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Strength drop attributed to grain growth", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Connection pool exhaustion", "Garbage collection pauses", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Upstream service slowdown", "CPU throttling by cgroup quota", "Disk IO contention"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Garbage collection pauses", "Upstream service slowdown", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Disk IO contention"], "domain": "software_ml", "evidence": "Engineer's handoff: p99 latency spikes every 56 s to 804 ms. GC logs show pauses <39 ms. DNS TTL for upstream is 56 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 2, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [56.0, 804.0, 39.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Connection pool exhaustion", "Garbage collection pauses", "Upstream service slowdown", "CPU throttling by cgroup quota", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Disk IO contention"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Garbage collection pauses", "Connection pool exhaustion", "CPU throttling by cgroup quota", "Disk IO contention", "Upstream service slowdown"], "domain": "software_ml", "evidence": "Operator note: p99 latency spikes every 256 s to 1452 ms. GC logs show pauses <18 ms. DNS TTL for upstream is 256 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. badge reader log entry unrelated Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 4, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [256.0, 1452.0, 18.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Garbage collection pauses", "Upstream service slowdown", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Disk IO contention"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "Connection pool exhaustion", "Disk IO contention", "CPU throttling by cgroup quota", "Garbage collection pauses"], "domain": "software_ml", "evidence": "Operator note: p99 latency spikes every 266 s to 928 ms. GC logs show pauses <14 ms. DNS TTL for upstream is 266 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'cpu throttling by cgroup quota' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [266.0, 928.0, 14.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Garbage collection pauses", "Disk IO contention", "Upstream service slowdown", "CPU throttling by cgroup quota", "Connection pool exhaustion"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Garbage collection pauses", "Upstream service slowdown", "CPU throttling by cgroup quota", "Disk IO contention", "Connection pool exhaustion"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): p99 latency spikes every 80 s to 912 ms. GC logs show pauses <10 ms. DNS TTL for upstream is 80 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [80.0, 912.0, 10.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Disk IO contention", "Garbage collection pauses", "Upstream service slowdown"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "Disk IO contention", "Connection pool exhaustion", "CPU throttling by cgroup quota", "Garbage collection pauses"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): p99 latency spikes every 72 s to 645 ms. GC logs show pauses <38 ms. DNS TTL for upstream is 72 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. Firmware banner printed at boot Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [72.0, 645.0, 38.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Garbage collection pauses", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Disk IO contention", "Connection pool exhaustion", "Upstream service slowdown", "CPU throttling by cgroup quota"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Garbage collection pauses", "CPU throttling by cgroup quota", "Upstream service slowdown", "Disk IO contention"], "domain": "software_ml", "evidence": "Field report: p99 latency spikes every 127 s to 988 ms. GC logs show pauses <31 ms. DNS TTL for upstream is 127 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'upstream service slowdown' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 1, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [127.0, 988.0, 31.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Upstream service slowdown", "Garbage collection pauses", "Connection pool exhaustion", "Disk IO contention", "CPU throttling by cgroup quota", "DNS re-resolution on TTL expiry with no resolver cache stalling requests"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Upstream service slowdown", "CPU throttling by cgroup quota", "Disk IO contention", "Garbage collection pauses"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): p99 latency spikes every 258 s to 623 ms. GC logs show pauses <16 ms. DNS TTL for upstream is 258 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 5, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [258.0, 623.0, 16.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Disk IO contention", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Upstream service slowdown", "CPU throttling by cgroup quota", "Garbage collection pauses", "Connection pool exhaustion"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Garbage collection pauses", "Disk IO contention", "Upstream service slowdown", "CPU throttling by cgroup quota"], "domain": "software_ml", "evidence": "Field report: p99 latency spikes every 235 s to 502 ms. GC logs show pauses <27 ms. DNS TTL for upstream is 235 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. duplicate log line suppressed x12 [INFO] heartbeat ok", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 1, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [235.0, 502.0, 27.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Garbage collection pauses", "Upstream service slowdown", "Connection pool exhaustion", "CPU throttling by cgroup quota", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Disk IO contention"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "Disk IO contention", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Garbage collection pauses"], "domain": "software_ml", "evidence": "Ticket summary: p99 latency spikes every 292 s to 1221 ms. GC logs show pauses <28 ms. DNS TTL for upstream is 292 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'garbage collection pauses' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 4, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [292.0, 1221.0, 28.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Connection pool exhaustion", "Upstream service slowdown", "Garbage collection pauses", "Disk IO contention", "CPU throttling by cgroup quota"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Disk IO contention", "Upstream service slowdown", "Garbage collection pauses", "CPU throttling by cgroup quota"], "domain": "software_ml", "evidence": "Operator note: p99 latency spikes every 161 s to 878 ms. GC logs show pauses <38 ms. DNS TTL for upstream is 161 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [161.0, 878.0, 38.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["CPU throttling by cgroup quota", "Upstream service slowdown", "Garbage collection pauses", "Disk IO contention", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Connection pool exhaustion"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Upstream service slowdown", "CPU throttling by cgroup quota", "Disk IO contention", "Garbage collection pauses"], "domain": "software_ml", "evidence": "p99 latency spikes every 32 s to 1686 ms. GC logs show pauses <30 ms. DNS TTL for upstream is 32 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. Note: shift change at 14:00 badge reader log entry unrelated", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 4, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [32.0, 1686.0, 30.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Disk IO contention", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Upstream service slowdown", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Garbage collection pauses"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "CPU throttling by cgroup quota", "Garbage collection pauses", "Connection pool exhaustion", "Disk IO contention"], "domain": "software_ml", "evidence": "Operator note: p99 latency spikes every 262 s to 558 ms. GC logs show pauses <37 ms. DNS TTL for upstream is 262 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'garbage collection pauses' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 4, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [262.0, 558.0, 37.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Disk IO contention", "Garbage collection pauses", "CPU throttling by cgroup quota", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Connection pool exhaustion", "Upstream service slowdown"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Garbage collection pauses", "Disk IO contention", "Upstream service slowdown", "CPU throttling by cgroup quota"], "domain": "software_ml", "evidence": "p99 latency spikes every 39 s to 1971 ms. GC logs show pauses <17 ms. DNS TTL for upstream is 39 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 3, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [39.0, 1971.0, 17.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Garbage collection pauses", "Upstream service slowdown", "Disk IO contention", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "CPU throttling by cgroup quota", "Connection pool exhaustion"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "Connection pool exhaustion", "Garbage collection pauses", "CPU throttling by cgroup quota", "Disk IO contention"], "domain": "software_ml", "evidence": "p99 latency spikes every 55 s to 1409 ms. GC logs show pauses <26 ms. DNS TTL for upstream is 55 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. HMI screensaver restarted Note: shift change at 14:00", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 3, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [55.0, 1409.0, 26.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Upstream service slowdown", "Connection pool exhaustion", "CPU throttling by cgroup quota", "Disk IO contention", "Garbage collection pauses"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Garbage collection pauses", "Disk IO contention"], "domain": "software_ml", "evidence": "Engineer's handoff: p99 latency spikes every 245 s to 1690 ms. GC logs show pauses <24 ms. DNS TTL for upstream is 245 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'cpu throttling by cgroup quota' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [245.0, 1690.0, 24.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Garbage collection pauses", "Upstream service slowdown", "Disk IO contention", "Connection pool exhaustion", "CPU throttling by cgroup quota"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Garbage collection pauses", "Disk IO contention", "Upstream service slowdown", "CPU throttling by cgroup quota"], "domain": "software_ml", "evidence": "Field report: p99 latency spikes every 74 s to 1659 ms. GC logs show pauses <40 ms. DNS TTL for upstream is 74 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [74.0, 1659.0, 40.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Garbage collection pauses", "Upstream service slowdown", "CPU throttling by cgroup quota", "Connection pool exhaustion", "Disk IO contention", "DNS re-resolution on TTL expiry with no resolver cache stalling requests"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Disk IO contention", "Garbage collection pauses", "CPU throttling by cgroup quota", "Upstream service slowdown"], "domain": "software_ml", "evidence": "Operator note: p99 latency spikes every 149 s to 1305 ms. GC logs show pauses <16 ms. DNS TTL for upstream is 149 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. duplicate log line suppressed x12 [INFO] heartbeat ok", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 5, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [149.0, 1305.0, 16.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Disk IO contention", "Garbage collection pauses", "Upstream service slowdown", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "CPU throttling by cgroup quota", "Connection pool exhaustion"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["CPU throttling by cgroup quota", "Upstream service slowdown", "Garbage collection pauses", "Disk IO contention", "Connection pool exhaustion"], "domain": "software_ml", "evidence": "Engineer's handoff: p99 latency spikes every 189 s to 1132 ms. GC logs show pauses <26 ms. DNS TTL for upstream is 189 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'garbage collection pauses' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 3, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [189.0, 1132.0, 26.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Upstream service slowdown", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Connection pool exhaustion", "Garbage collection pauses", "Disk IO contention", "CPU throttling by cgroup quota"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Connection pool exhaustion", "Upstream service slowdown", "CPU throttling by cgroup quota", "Garbage collection pauses", "Disk IO contention"], "domain": "software_ml", "evidence": "p99 latency spikes every 48 s to 1870 ms. GC logs show pauses <13 ms. DNS TTL for upstream is 48 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 1, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [48.0, 1870.0, 13.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Disk IO contention", "Upstream service slowdown", "Garbage collection pauses", "CPU throttling by cgroup quota", "Connection pool exhaustion"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["CPU throttling by cgroup quota", "Disk IO contention", "Upstream service slowdown", "Connection pool exhaustion", "Garbage collection pauses"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): p99 latency spikes every 191 s to 1490 ms. GC logs show pauses <27 ms. DNS TTL for upstream is 191 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. Note: shift change at 14:00 [INFO] heartbeat ok", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 0, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [191.0, 1490.0, 27.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Disk IO contention", "Connection pool exhaustion", "Upstream service slowdown", "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "Garbage collection pauses", "CPU throttling by cgroup quota"], "difficulty": "hard", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Upstream service slowdown", "Connection pool exhaustion", "Garbage collection pauses", "Disk IO contention", "CPU throttling by cgroup quota"], "domain": "software_ml", "evidence": "Engineer's handoff: p99 latency spikes every 237 s to 476 ms. GC logs show pauses <27 ms. DNS TTL for upstream is 237 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace. A colleague suspects 'garbage collection pauses' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 3, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [237.0, 476.0, 27.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Garbage collection pauses", "CPU throttling by cgroup quota", "Upstream service slowdown", "Disk IO contention", "Connection pool exhaustion", "DNS re-resolution on TTL expiry with no resolver cache stalling requests"], "difficulty": "medium", "discriminating_followups": ["Does enabling a DNS cache remove spikes?", "Is spike period equal to TTL when TTL is changed?", "Do spans show time in DNS resolution?"], "distractors": ["Disk IO contention", "Connection pool exhaustion", "Garbage collection pauses", "CPU throttling by cgroup quota", "Upstream service slowdown"], "domain": "software_ml", "evidence": "Field report: p99 latency spikes every 246 s to 1868 ms. GC logs show pauses <30 ms. DNS TTL for upstream is 246 s; resolver cache disabled in container image. Spikes coincide with DNS lookups in the trace.", "failure_category": "instrumentation_artifact", "gold_cause": "DNS re-resolution on TTL expiry with no resolver cache stalling requests", "gold_hypothesis_index": 5, "group_id": "fam:SW-GC-VS-DNS", "id": "SYN-SW-GC-VS-DNS-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Spike period matching DNS TTL, and GC pauses far below spike size, points away from GC."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "backend", "tabular_stats": [246.0, 1868.0, 30.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Periodic latency spikes", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Job really is idle", "Clock skew between nodes", "Exporter crash", "cgroup accounting bug", "Dashboard query typo"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Clock skew between nodes", "cgroup accounting bug", "Dashboard query typo", "Exporter crash"], "domain": "software_ml", "evidence": "Field report: Batch job CPU bursts of 6 s visible in node logs but absent in Prometheus graphs. Scrape interval 56 s; rate() window 227 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 0, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [6.0, 56.0, 227.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Dashboard query typo", "cgroup accounting bug", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Job really is idle", "Clock skew between nodes", "Exporter crash"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Clock skew between nodes", "cgroup accounting bug", "Dashboard query typo", "Exporter crash"], "domain": "software_ml", "evidence": "Ticket summary: Batch job CPU bursts of 4 s visible in node logs but absent in Prometheus graphs. Scrape interval 37 s; rate() window 273 s. Users conclude CPU usage is low. Note: shift change at 14:00 Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [4.0, 37.0, 273.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "Job really is idle", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Dashboard query typo", "Exporter crash", "cgroup accounting bug"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Dashboard query typo", "cgroup accounting bug", "Clock skew between nodes", "Exporter crash", "Job really is idle"], "domain": "software_ml", "evidence": "Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 41 s; rate() window 122 s. Users conclude CPU usage is low. A colleague suspects 'job really is idle' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 41.0, 122.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Exporter crash", "Dashboard query typo", "cgroup accounting bug", "Job really is idle", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Clock skew between nodes"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Dashboard query typo", "Exporter crash", "Clock skew between nodes", "cgroup accounting bug"], "domain": "software_ml", "evidence": "Field report: Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 50 s; rate() window 235 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 4, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 50.0, 235.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Exporter crash", "Dashboard query typo", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Job really is idle", "Clock skew between nodes", "cgroup accounting bug"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Exporter crash", "Dashboard query typo", "Job really is idle", "cgroup accounting bug", "Clock skew between nodes"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): Batch job CPU bursts of 3 s visible in node logs but absent in Prometheus graphs. Scrape interval 59 s; rate() window 142 s. Users conclude CPU usage is low. Firmware banner printed at boot duplicate log line suppressed x12", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [3.0, 59.0, 142.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "cgroup accounting bug", "Dashboard query typo", "Job really is idle"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Clock skew between nodes", "Exporter crash", "cgroup accounting bug", "Job really is idle", "Dashboard query typo"], "domain": "software_ml", "evidence": "Operator note: Batch job CPU bursts of 6 s visible in node logs but absent in Prometheus graphs. Scrape interval 43 s; rate() window 217 s. Users conclude CPU usage is low. A colleague suspects 'job really is idle' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [6.0, 43.0, 217.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Dashboard query typo", "cgroup accounting bug", "Clock skew between nodes", "Job really is idle"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Clock skew between nodes", "cgroup accounting bug", "Job really is idle", "Dashboard query typo", "Exporter crash"], "domain": "software_ml", "evidence": "Ticket summary: Batch job CPU bursts of 3 s visible in node logs but absent in Prometheus graphs. Scrape interval 47 s; rate() window 132 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 1, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [3.0, 47.0, 132.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["cgroup accounting bug", "Dashboard query typo", "Clock skew between nodes", "Job really is idle", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Exporter crash"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Dashboard query typo", "Clock skew between nodes", "Job really is idle", "Exporter crash", "cgroup accounting bug"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): Batch job CPU bursts of 5 s visible in node logs but absent in Prometheus graphs. Scrape interval 36 s; rate() window 268 s. Users conclude CPU usage is low. badge reader log entry unrelated Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 4, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [5.0, 36.0, 268.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Dashboard query typo", "Job really is idle", "cgroup accounting bug", "Clock skew between nodes", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Exporter crash", "cgroup accounting bug", "Dashboard query typo", "Clock skew between nodes"], "domain": "software_ml", "evidence": "Engineer's handoff: Batch job CPU bursts of 6 s visible in node logs but absent in Prometheus graphs. Scrape interval 39 s; rate() window 201 s. Users conclude CPU usage is low. A colleague suspects 'clock skew between nodes' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 5, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [6.0, 39.0, 201.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Dashboard query typo", "cgroup accounting bug", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Exporter crash", "Clock skew between nodes", "Job really is idle"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Clock skew between nodes", "Dashboard query typo", "Exporter crash", "cgroup accounting bug"], "domain": "software_ml", "evidence": "Engineer's handoff: Batch job CPU bursts of 4 s visible in node logs but absent in Prometheus graphs. Scrape interval 52 s; rate() window 196 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [4.0, 52.0, 196.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "Exporter crash", "Job really is idle", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "cgroup accounting bug", "Dashboard query typo"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["cgroup accounting bug", "Clock skew between nodes", "Job really is idle", "Exporter crash", "Dashboard query typo"], "domain": "software_ml", "evidence": "Ticket summary: Batch job CPU bursts of 5 s visible in node logs but absent in Prometheus graphs. Scrape interval 57 s; rate() window 215 s. Users conclude CPU usage is low. Unrelated: cafeteria chiller alarm acknowledged [INFO] heartbeat ok", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 3, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [5.0, 57.0, 215.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "cgroup accounting bug", "Job really is idle", "Dashboard query typo"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Dashboard query typo", "Clock skew between nodes", "Exporter crash", "cgroup accounting bug"], "domain": "software_ml", "evidence": "Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 35 s; rate() window 213 s. Users conclude CPU usage is low. A colleague suspects 'clock skew between nodes' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 35.0, 213.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Dashboard query typo", "Clock skew between nodes", "Job really is idle", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "cgroup accounting bug"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Clock skew between nodes", "cgroup accounting bug", "Exporter crash", "Dashboard query typo", "Job really is idle"], "domain": "software_ml", "evidence": "Shift log excerpt (cleansed): Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 32 s; rate() window 184 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 4, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 32.0, 184.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["cgroup accounting bug", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Clock skew between nodes", "Dashboard query typo", "Job really is idle"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Clock skew between nodes", "Job really is idle", "Exporter crash", "cgroup accounting bug", "Dashboard query typo"], "domain": "software_ml", "evidence": "Ticket summary: Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 35 s; rate() window 179 s. Users conclude CPU usage is low. Unrelated: cafeteria chiller alarm acknowledged [INFO] heartbeat ok", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 35.0, 179.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Job really is idle", "cgroup accounting bug", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Exporter crash", "Clock skew between nodes", "Dashboard query typo"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Dashboard query typo", "Clock skew between nodes", "cgroup accounting bug", "Exporter crash"], "domain": "software_ml", "evidence": "Engineer's handoff: Batch job CPU bursts of 4 s visible in node logs but absent in Prometheus graphs. Scrape interval 35 s; rate() window 241 s. Users conclude CPU usage is low. A colleague suspects 'exporter crash' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 2, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [4.0, 35.0, 241.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Dashboard query typo", "cgroup accounting bug", "Job really is idle", "Exporter crash", "Clock skew between nodes", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Job really is idle", "Clock skew between nodes", "cgroup accounting bug", "Dashboard query typo", "Exporter crash"], "domain": "software_ml", "evidence": "Batch job CPU bursts of 6 s visible in node logs but absent in Prometheus graphs. Scrape interval 42 s; rate() window 256 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 5, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [6.0, 42.0, 256.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Exporter crash", "cgroup accounting bug", "Dashboard query typo", "Job really is idle"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Dashboard query typo", "cgroup accounting bug", "Job really is idle", "Exporter crash", "Clock skew between nodes"], "domain": "software_ml", "evidence": "Field report: Batch job CPU bursts of 8 s visible in node logs but absent in Prometheus graphs. Scrape interval 46 s; rate() window 116 s. Users conclude CPU usage is low. duplicate log line suppressed x12 HMI screensaver restarted", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 1, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [8.0, 46.0, 116.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "cgroup accounting bug", "Job really is idle", "Dashboard query typo", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["cgroup accounting bug", "Dashboard query typo", "Job really is idle", "Clock skew between nodes", "Exporter crash"], "domain": "software_ml", "evidence": "Operator note: Batch job CPU bursts of 4 s visible in node logs but absent in Prometheus graphs. Scrape interval 40 s; rate() window 159 s. Users conclude CPU usage is low. A colleague suspects 'exporter crash' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 5, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [4.0, 40.0, 159.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Exporter crash", "cgroup accounting bug", "Job really is idle", "Clock skew between nodes", "Dashboard query typo", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["cgroup accounting bug", "Dashboard query typo", "Exporter crash", "Clock skew between nodes", "Job really is idle"], "domain": "software_ml", "evidence": "Ticket summary: Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 36 s; rate() window 243 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 5, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 36.0, 243.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Exporter crash", "Job really is idle", "Dashboard query typo", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "cgroup accounting bug", "Clock skew between nodes"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Dashboard query typo", "Exporter crash", "Job really is idle", "cgroup accounting bug", "Clock skew between nodes"], "domain": "software_ml", "evidence": "Operator note: Batch job CPU bursts of 7 s visible in node logs but absent in Prometheus graphs. Scrape interval 37 s; rate() window 197 s. Users conclude CPU usage is low. badge reader log entry unrelated Note: shift change at 14:00", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 3, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [7.0, 37.0, 197.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "Clock skew between nodes", "Job really is idle", "cgroup accounting bug", "Dashboard query typo", "Exporter crash"], "difficulty": "hard", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Dashboard query typo", "cgroup accounting bug", "Clock skew between nodes", "Job really is idle", "Exporter crash"], "domain": "software_ml", "evidence": "Operator note: Batch job CPU bursts of 8 s visible in node logs but absent in Prometheus graphs. Scrape interval 31 s; rate() window 84 s. Users conclude CPU usage is low. A colleague suspects 'exporter crash' because a similar symptom was seen once before (no measurement attached).", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 0, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [8.0, 31.0, 84.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Clock skew between nodes", "cgroup accounting bug", "Dashboard query typo", "Job really is idle", "Exporter crash", "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows"], "difficulty": "medium", "discriminating_followups": ["Does a 1 s scrape reveal the bursts?", "Does max_over_time show them?", "Do node-local counters integrate the burst CPU seconds?"], "distractors": ["Clock skew between nodes", "Job really is idle", "Dashboard query typo", "Exporter crash", "cgroup accounting bug"], "domain": "software_ml", "evidence": "Operator note: Batch job CPU bursts of 6 s visible in node logs but absent in Prometheus graphs. Scrape interval 38 s; rate() window 249 s. Users conclude CPU usage is low.", "failure_category": "instrumentation_artifact", "gold_cause": "Scrape/aggregation aliasing: short bursts averaged out by long scrape and rate windows", "gold_hypothesis_index": 5, "group_id": "fam:SW-PROM-ALIAS", "id": "SYN-SW-PROM-ALIAS-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Sampling interval >> burst duration hides bursts (Nyquist)."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "observability", "tabular_stats": [6.0, 38.0, 249.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Missing CPU bursts in dashboards", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Impeller imbalance", "Abrasive solids erosion", "Bearing failure noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Discharge recirculation at low flow", "Motor electrical noise"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Bearing failure noise", "Motor electrical noise", "Discharge recirculation at low flow", "Abrasive solids erosion", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.4 m vs NPSHr 4.8 m. Suction valve found 53% throttled by operator to control flow. Fluid temp 70 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.8, 53.0, 70.0, 3.36, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Motor electrical noise", "Bearing failure noise", "Impeller imbalance", "Cavitation from throttled suction valve dropping NPSH available below required", "Discharge recirculation at low flow", "Abrasive solids erosion"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Motor electrical noise", "Abrasive solids erosion", "Discharge recirculation at low flow", "Bearing failure noise", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Operator note: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.2 m vs NPSHr 3.1 m. Suction valve found 57% throttled by operator to control flow. Fluid temp 60 C. Note: shift change at 14:00 badge reader log entry unrelated", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.1, 57.0, 60.0, 2.17, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Bearing failure noise", "Discharge recirculation at low flow", "Abrasive solids erosion", "Cavitation from throttled suction valve dropping NPSH available below required", "Impeller imbalance", "Motor electrical noise"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Motor electrical noise", "Impeller imbalance", "Abrasive solids erosion", "Bearing failure noise", "Discharge recirculation at low flow"], "domain": "fluids_thermal", "evidence": "Shift log excerpt (cleansed): Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.5 m vs NPSHr 3.6 m. Suction valve found 61% throttled by operator to control flow. Fluid temp 44 C. A colleague suspects 'bearing failure noise' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.6, 61.0, 44.0, 2.52, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Motor electrical noise", "Impeller imbalance", "Discharge recirculation at low flow", "Bearing failure noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Abrasive solids erosion"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Bearing failure noise", "Impeller imbalance", "Abrasive solids erosion", "Motor electrical noise", "Discharge recirculation at low flow"], "domain": "fluids_thermal", "evidence": "Shift log excerpt (cleansed): Pump crackling noise; impeller pitting at blade inlet. NPSHa 4.1 m vs NPSHr 5.8 m. Suction valve found 58% throttled by operator to control flow. Fluid temp 63 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 4, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [5.8, 58.0, 63.0, 4.06, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Discharge recirculation at low flow", "Bearing failure noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Impeller imbalance", "Motor electrical noise", "Abrasive solids erosion"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Bearing failure noise", "Discharge recirculation at low flow", "Abrasive solids erosion", "Motor electrical noise", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.6 m vs NPSHr 3.7 m. Suction valve found 41% throttled by operator to control flow. Fluid temp 55 C. [INFO] heartbeat ok Firmware banner printed at boot", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 2, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.7, 41.0, 55.0, 2.59, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Abrasive solids erosion", "Impeller imbalance", "Discharge recirculation at low flow", "Bearing failure noise", "Motor electrical noise", "Cavitation from throttled suction valve dropping NPSH available below required"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Bearing failure noise", "Discharge recirculation at low flow", "Impeller imbalance", "Abrasive solids erosion", "Motor electrical noise"], "domain": "fluids_thermal", "evidence": "Engineer's handoff: Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.3 m vs NPSHr 4.7 m. Suction valve found 52% throttled by operator to control flow. Fluid temp 75 C. A colleague suspects 'bearing failure noise' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 5, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.7, 52.0, 75.0, 3.29, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Impeller imbalance", "Discharge recirculation at low flow", "Cavitation from throttled suction valve dropping NPSH available below required", "Bearing failure noise", "Abrasive solids erosion", "Motor electrical noise"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Abrasive solids erosion", "Bearing failure noise", "Motor electrical noise", "Impeller imbalance", "Discharge recirculation at low flow"], "domain": "fluids_thermal", "evidence": "Ticket summary: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.4 m vs NPSHr 3.5 m. Suction valve found 68% throttled by operator to control flow. Fluid temp 48 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 2, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.5, 68.0, 48.0, 2.4499999999999997, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Cavitation from throttled suction valve dropping NPSH available below required", "Bearing failure noise", "Discharge recirculation at low flow", "Motor electrical noise", "Abrasive solids erosion", "Impeller imbalance"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Bearing failure noise", "Discharge recirculation at low flow", "Abrasive solids erosion", "Motor electrical noise", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.4 m vs NPSHr 4.8 m. Suction valve found 44% throttled by operator to control flow. Fluid temp 55 C. duplicate log line suppressed x12 Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 0, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.8, 44.0, 55.0, 3.36, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Impeller imbalance", "Discharge recirculation at low flow", "Bearing failure noise", "Motor electrical noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Abrasive solids erosion"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Motor electrical noise", "Abrasive solids erosion", "Discharge recirculation at low flow", "Impeller imbalance", "Bearing failure noise"], "domain": "fluids_thermal", "evidence": "Operator note: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.9 m vs NPSHr 4.1 m. Suction valve found 62% throttled by operator to control flow. Fluid temp 55 C. A colleague suspects 'abrasive solids erosion' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 4, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.1, 62.0, 55.0, 2.8699999999999997, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Abrasive solids erosion", "Motor electrical noise", "Discharge recirculation at low flow", "Impeller imbalance", "Cavitation from throttled suction valve dropping NPSH available below required", "Bearing failure noise"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Bearing failure noise", "Motor electrical noise", "Impeller imbalance", "Discharge recirculation at low flow", "Abrasive solids erosion"], "domain": "fluids_thermal", "evidence": "Ticket summary: Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.2 m vs NPSHr 4.6 m. Suction valve found 63% throttled by operator to control flow. Fluid temp 73 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 4, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.6, 63.0, 73.0, 3.2199999999999998, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Impeller imbalance", "Abrasive solids erosion", "Bearing failure noise", "Discharge recirculation at low flow", "Cavitation from throttled suction valve dropping NPSH available below required", "Motor electrical noise"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Discharge recirculation at low flow", "Abrasive solids erosion", "Impeller imbalance", "Motor electrical noise", "Bearing failure noise"], "domain": "fluids_thermal", "evidence": "Field report: Pump crackling noise; impeller pitting at blade inlet. NPSHa 4.0 m vs NPSHr 5.7 m. Suction valve found 56% throttled by operator to control flow. Fluid temp 51 C. Firmware banner printed at boot badge reader log entry unrelated", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 4, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [5.7, 56.0, 51.0, 3.9899999999999998, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Impeller imbalance", "Motor electrical noise", "Bearing failure noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Abrasive solids erosion", "Discharge recirculation at low flow"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Discharge recirculation at low flow", "Abrasive solids erosion", "Motor electrical noise", "Impeller imbalance", "Bearing failure noise"], "domain": "fluids_thermal", "evidence": "Engineer's handoff: Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.1 m vs NPSHr 4.4 m. Suction valve found 64% throttled by operator to control flow. Fluid temp 66 C. A colleague suspects 'impeller imbalance' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.4, 64.0, 66.0, 3.08, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Motor electrical noise", "Impeller imbalance", "Bearing failure noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Abrasive solids erosion", "Discharge recirculation at low flow"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Abrasive solids erosion", "Bearing failure noise", "Motor electrical noise", "Impeller imbalance", "Discharge recirculation at low flow"], "domain": "fluids_thermal", "evidence": "Ticket summary: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.8 m vs NPSHr 4.0 m. Suction valve found 64% throttled by operator to control flow. Fluid temp 70 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.0, 64.0, 70.0, 2.8, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Abrasive solids erosion", "Cavitation from throttled suction valve dropping NPSH available below required", "Motor electrical noise", "Discharge recirculation at low flow", "Impeller imbalance", "Bearing failure noise"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Motor electrical noise", "Bearing failure noise", "Abrasive solids erosion", "Discharge recirculation at low flow", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Engineer's handoff: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.7 m vs NPSHr 3.8 m. Suction valve found 69% throttled by operator to control flow. Fluid temp 67 C. duplicate log line suppressed x12 Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 1, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.8, 69.0, 67.0, 2.6599999999999997, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Bearing failure noise", "Abrasive solids erosion", "Motor electrical noise", "Discharge recirculation at low flow", "Impeller imbalance", "Cavitation from throttled suction valve dropping NPSH available below required"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Impeller imbalance", "Bearing failure noise", "Motor electrical noise", "Abrasive solids erosion", "Discharge recirculation at low flow"], "domain": "fluids_thermal", "evidence": "Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.2 m vs NPSHr 3.2 m. Suction valve found 50% throttled by operator to control flow. Fluid temp 64 C. A colleague suspects 'bearing failure noise' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 5, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.2, 50.0, 64.0, 2.2399999999999998, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Motor electrical noise", "Bearing failure noise", "Abrasive solids erosion", "Cavitation from throttled suction valve dropping NPSH available below required", "Discharge recirculation at low flow", "Impeller imbalance"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Motor electrical noise", "Discharge recirculation at low flow", "Impeller imbalance", "Bearing failure noise", "Abrasive solids erosion"], "domain": "fluids_thermal", "evidence": "Operator note: Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.3 m vs NPSHr 4.7 m. Suction valve found 45% throttled by operator to control flow. Fluid temp 66 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.7, 45.0, 66.0, 3.29, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Abrasive solids erosion", "Motor electrical noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Impeller imbalance", "Discharge recirculation at low flow", "Bearing failure noise"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Discharge recirculation at low flow", "Motor electrical noise", "Bearing failure noise", "Abrasive solids erosion", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Shift log excerpt (cleansed): Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.5 m vs NPSHr 5.0 m. Suction valve found 66% throttled by operator to control flow. Fluid temp 54 C. HMI screensaver restarted badge reader log entry unrelated", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 2, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [5.0, 66.0, 54.0, 3.5, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Cavitation from throttled suction valve dropping NPSH available below required", "Motor electrical noise", "Discharge recirculation at low flow", "Bearing failure noise", "Impeller imbalance", "Abrasive solids erosion"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Impeller imbalance", "Discharge recirculation at low flow", "Bearing failure noise", "Abrasive solids erosion", "Motor electrical noise"], "domain": "fluids_thermal", "evidence": "Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.6 m vs NPSHr 5.2 m. Suction valve found 56% throttled by operator to control flow. Fluid temp 79 C. A colleague suspects 'bearing failure noise' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 0, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [5.2, 56.0, 79.0, 3.6399999999999997, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Bearing failure noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Abrasive solids erosion", "Impeller imbalance", "Discharge recirculation at low flow", "Motor electrical noise"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Impeller imbalance", "Abrasive solids erosion", "Motor electrical noise", "Discharge recirculation at low flow", "Bearing failure noise"], "domain": "fluids_thermal", "evidence": "Operator note: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.9 m vs NPSHr 4.1 m. Suction valve found 53% throttled by operator to control flow. Fluid temp 49 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 1, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.1, 53.0, 49.0, 2.8699999999999997, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Discharge recirculation at low flow", "Abrasive solids erosion", "Bearing failure noise", "Motor electrical noise", "Cavitation from throttled suction valve dropping NPSH available below required", "Impeller imbalance"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Abrasive solids erosion", "Bearing failure noise", "Motor electrical noise", "Impeller imbalance", "Discharge recirculation at low flow"], "domain": "fluids_thermal", "evidence": "Field report: Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.8 m vs NPSHr 5.5 m. Suction valve found 42% throttled by operator to control flow. Fluid temp 63 C. HMI screensaver restarted badge reader log entry unrelated", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 4, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [5.5, 42.0, 63.0, 3.8499999999999996, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Bearing failure noise", "Impeller imbalance", "Abrasive solids erosion", "Cavitation from throttled suction valve dropping NPSH available below required", "Motor electrical noise", "Discharge recirculation at low flow"], "difficulty": "hard", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Abrasive solids erosion", "Motor electrical noise", "Discharge recirculation at low flow", "Impeller imbalance", "Bearing failure noise"], "domain": "fluids_thermal", "evidence": "Engineer's handoff: Pump crackling noise; impeller pitting at blade inlet. NPSHa 2.7 m vs NPSHr 3.9 m. Suction valve found 65% throttled by operator to control flow. Fluid temp 74 C. A colleague suspects 'impeller imbalance' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 3, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [3.9, 65.0, 74.0, 2.73, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Cavitation from throttled suction valve dropping NPSH available below required", "Abrasive solids erosion", "Bearing failure noise", "Motor electrical noise", "Discharge recirculation at low flow", "Impeller imbalance"], "difficulty": "medium", "discriminating_followups": ["Does opening suction and throttling discharge instead stop noise?", "Pitting location inlet vs outlet?", "Suction pressure vs vapor pressure at temp?"], "distractors": ["Abrasive solids erosion", "Bearing failure noise", "Discharge recirculation at low flow", "Motor electrical noise", "Impeller imbalance"], "domain": "fluids_thermal", "evidence": "Shift log excerpt (cleansed): Pump crackling noise; impeller pitting at blade inlet. NPSHa 3.1 m vs NPSHr 4.5 m. Suction valve found 68% throttled by operator to control flow. Fluid temp 51 C.", "failure_category": "protocol_human_loophole", "gold_cause": "Cavitation from throttled suction valve dropping NPSH available below required", "gold_hypothesis_index": 0, "group_id": "fam:FLU-CAVITATION", "id": "SYN-FLU-CAVITATION-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["NPSHa < NPSHr -> vapor bubbles collapse at blade inlet causing pitting."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "pumps", "tabular_stats": [4.5, 68.0, 51.0, 3.15, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Pump noise and impeller erosion", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Cell line passage drift", "Plastics leachables", "ELISA cross-reactivity", "Incubator CO2 failure", "Mycoplasma contamination", "Endotoxin-contaminated serum lot activating TLR4"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Cell line passage drift", "Plastics leachables", "Mycoplasma contamination", "ELISA cross-reactivity", "Incubator CO2 failure"], "domain": "biology_lab", "evidence": "Field report: Unstimulated macrophages secrete TNF 1069 pg/mL (normally <50). Started with new FBS lot 7660; endotoxin 3.7 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 5, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1069.0, 7660.0, 3.7, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Plastics leachables", "Mycoplasma contamination", "Incubator CO2 failure", "Cell line passage drift", "Endotoxin-contaminated serum lot activating TLR4", "ELISA cross-reactivity"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["ELISA cross-reactivity", "Plastics leachables", "Cell line passage drift", "Mycoplasma contamination", "Incubator CO2 failure"], "domain": "biology_lab", "evidence": "Field report: Unstimulated macrophages secrete TNF 1420 pg/mL (normally <50). Started with new FBS lot 2388; endotoxin 3.4 EU/mL by LAL. Old lot cultures normal. [INFO] heartbeat ok HMI screensaver restarted", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 4, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1420.0, 2388.0, 3.4, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["ELISA cross-reactivity", "Mycoplasma contamination", "Endotoxin-contaminated serum lot activating TLR4", "Cell line passage drift", "Plastics leachables", "Incubator CO2 failure"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Incubator CO2 failure", "Mycoplasma contamination", "Cell line passage drift", "Plastics leachables", "ELISA cross-reactivity"], "domain": "biology_lab", "evidence": "Ticket summary: Unstimulated macrophages secrete TNF 590 pg/mL (normally <50). Started with new FBS lot 8968; endotoxin 2.1 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'mycoplasma contamination' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [590.0, 8968.0, 2.1, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Mycoplasma contamination", "Incubator CO2 failure", "ELISA cross-reactivity", "Plastics leachables", "Endotoxin-contaminated serum lot activating TLR4", "Cell line passage drift"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Mycoplasma contamination", "ELISA cross-reactivity", "Cell line passage drift", "Plastics leachables", "Incubator CO2 failure"], "domain": "biology_lab", "evidence": "Field report: Unstimulated macrophages secrete TNF 1312 pg/mL (normally <50). Started with new FBS lot 6082; endotoxin 3.6 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 4, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1312.0, 6082.0, 3.6, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["ELISA cross-reactivity", "Incubator CO2 failure", "Mycoplasma contamination", "Endotoxin-contaminated serum lot activating TLR4", "Cell line passage drift", "Plastics leachables"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Cell line passage drift", "ELISA cross-reactivity", "Plastics leachables", "Mycoplasma contamination", "Incubator CO2 failure"], "domain": "biology_lab", "evidence": "Field report: Unstimulated macrophages secrete TNF 964 pg/mL (normally <50). Started with new FBS lot 6692; endotoxin 1.3 EU/mL by LAL. Old lot cultures normal. badge reader log entry unrelated Note: shift change at 14:00", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 3, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [964.0, 6692.0, 1.3, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Plastics leachables", "Mycoplasma contamination", "ELISA cross-reactivity", "Incubator CO2 failure", "Endotoxin-contaminated serum lot activating TLR4", "Cell line passage drift"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Incubator CO2 failure", "Mycoplasma contamination", "ELISA cross-reactivity", "Plastics leachables", "Cell line passage drift"], "domain": "biology_lab", "evidence": "Operator note: Unstimulated macrophages secrete TNF 1423 pg/mL (normally <50). Started with new FBS lot 7014; endotoxin 2.0 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'elisa cross-reactivity' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 4, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1423.0, 7014.0, 2.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Mycoplasma contamination", "Incubator CO2 failure", "Cell line passage drift", "Endotoxin-contaminated serum lot activating TLR4", "Plastics leachables", "ELISA cross-reactivity"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Cell line passage drift", "Mycoplasma contamination", "Plastics leachables", "Incubator CO2 failure", "ELISA cross-reactivity"], "domain": "biology_lab", "evidence": "Engineer's handoff: Unstimulated macrophages secrete TNF 554 pg/mL (normally <50). Started with new FBS lot 3114; endotoxin 4.2 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 3, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [554.0, 3114.0, 4.2, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Cell line passage drift", "Plastics leachables", "Endotoxin-contaminated serum lot activating TLR4", "Mycoplasma contamination", "Incubator CO2 failure", "ELISA cross-reactivity"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["ELISA cross-reactivity", "Plastics leachables", "Cell line passage drift", "Incubator CO2 failure", "Mycoplasma contamination"], "domain": "biology_lab", "evidence": "Unstimulated macrophages secrete TNF 1242 pg/mL (normally <50). Started with new FBS lot 4433; endotoxin 3.9 EU/mL by LAL. Old lot cultures normal. Unrelated: cafeteria chiller alarm acknowledged duplicate log line suppressed x12", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1242.0, 4433.0, 3.9, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["ELISA cross-reactivity", "Incubator CO2 failure", "Plastics leachables", "Endotoxin-contaminated serum lot activating TLR4", "Mycoplasma contamination", "Cell line passage drift"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Incubator CO2 failure", "Plastics leachables", "Cell line passage drift", "Mycoplasma contamination", "ELISA cross-reactivity"], "domain": "biology_lab", "evidence": "Operator note: Unstimulated macrophages secrete TNF 308 pg/mL (normally <50). Started with new FBS lot 7477; endotoxin 4.8 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'elisa cross-reactivity' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 3, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [308.0, 7477.0, 4.8, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Cell line passage drift", "Mycoplasma contamination", "Incubator CO2 failure", "Plastics leachables", "Endotoxin-contaminated serum lot activating TLR4", "ELISA cross-reactivity"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Mycoplasma contamination", "ELISA cross-reactivity", "Plastics leachables", "Cell line passage drift", "Incubator CO2 failure"], "domain": "biology_lab", "evidence": "Operator note: Unstimulated macrophages secrete TNF 320 pg/mL (normally <50). Started with new FBS lot 1078; endotoxin 3.4 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 4, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [320.0, 1078.0, 3.4, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Endotoxin-contaminated serum lot activating TLR4", "Cell line passage drift", "Incubator CO2 failure", "ELISA cross-reactivity", "Plastics leachables", "Mycoplasma contamination"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Incubator CO2 failure", "ELISA cross-reactivity", "Plastics leachables", "Cell line passage drift", "Mycoplasma contamination"], "domain": "biology_lab", "evidence": "Ticket summary: Unstimulated macrophages secrete TNF 1444 pg/mL (normally <50). Started with new FBS lot 5650; endotoxin 4.9 EU/mL by LAL. Old lot cultures normal. HMI screensaver restarted Note: shift change at 14:00", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 0, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1444.0, 5650.0, 4.9, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Plastics leachables", "Endotoxin-contaminated serum lot activating TLR4", "Mycoplasma contamination", "Incubator CO2 failure", "ELISA cross-reactivity", "Cell line passage drift"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Incubator CO2 failure", "ELISA cross-reactivity", "Cell line passage drift", "Plastics leachables", "Mycoplasma contamination"], "domain": "biology_lab", "evidence": "Unstimulated macrophages secrete TNF 1375 pg/mL (normally <50). Started with new FBS lot 4561; endotoxin 4.9 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'cell line passage drift' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 1, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1375.0, 4561.0, 4.9, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Incubator CO2 failure", "Mycoplasma contamination", "Endotoxin-contaminated serum lot activating TLR4", "Plastics leachables", "ELISA cross-reactivity", "Cell line passage drift"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["ELISA cross-reactivity", "Mycoplasma contamination", "Plastics leachables", "Incubator CO2 failure", "Cell line passage drift"], "domain": "biology_lab", "evidence": "Ticket summary: Unstimulated macrophages secrete TNF 694 pg/mL (normally <50). Started with new FBS lot 7743; endotoxin 1.0 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [694.0, 7743.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Mycoplasma contamination", "ELISA cross-reactivity", "Endotoxin-contaminated serum lot activating TLR4", "Cell line passage drift", "Incubator CO2 failure", "Plastics leachables"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["ELISA cross-reactivity", "Cell line passage drift", "Mycoplasma contamination", "Plastics leachables", "Incubator CO2 failure"], "domain": "biology_lab", "evidence": "Shift log excerpt (cleansed): Unstimulated macrophages secrete TNF 572 pg/mL (normally <50). Started with new FBS lot 2136; endotoxin 2.3 EU/mL by LAL. Old lot cultures normal. duplicate log line suppressed x12 [INFO] heartbeat ok", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [572.0, 2136.0, 2.3, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Plastics leachables", "Incubator CO2 failure", "Endotoxin-contaminated serum lot activating TLR4", "Mycoplasma contamination", "ELISA cross-reactivity", "Cell line passage drift"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Mycoplasma contamination", "Incubator CO2 failure", "ELISA cross-reactivity", "Plastics leachables", "Cell line passage drift"], "domain": "biology_lab", "evidence": "Unstimulated macrophages secrete TNF 1497 pg/mL (normally <50). Started with new FBS lot 2230; endotoxin 3.7 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'cell line passage drift' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1497.0, 2230.0, 3.7, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Plastics leachables", "Mycoplasma contamination", "Cell line passage drift", "Endotoxin-contaminated serum lot activating TLR4", "Incubator CO2 failure", "ELISA cross-reactivity"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Plastics leachables", "Incubator CO2 failure", "Mycoplasma contamination", "ELISA cross-reactivity", "Cell line passage drift"], "domain": "biology_lab", "evidence": "Shift log excerpt (cleansed): Unstimulated macrophages secrete TNF 1399 pg/mL (normally <50). Started with new FBS lot 8600; endotoxin 1.8 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 3, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1399.0, 8600.0, 1.8, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Plastics leachables", "Mycoplasma contamination", "Endotoxin-contaminated serum lot activating TLR4", "ELISA cross-reactivity", "Incubator CO2 failure", "Cell line passage drift"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Plastics leachables", "Incubator CO2 failure", "Mycoplasma contamination", "Cell line passage drift", "ELISA cross-reactivity"], "domain": "biology_lab", "evidence": "Field report: Unstimulated macrophages secrete TNF 570 pg/mL (normally <50). Started with new FBS lot 8713; endotoxin 2.0 EU/mL by LAL. Old lot cultures normal. badge reader log entry unrelated duplicate log line suppressed x12", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [570.0, 8713.0, 2.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Endotoxin-contaminated serum lot activating TLR4", "Incubator CO2 failure", "Cell line passage drift", "Plastics leachables", "Mycoplasma contamination", "ELISA cross-reactivity"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Cell line passage drift", "Plastics leachables", "Incubator CO2 failure", "Mycoplasma contamination", "ELISA cross-reactivity"], "domain": "biology_lab", "evidence": "Field report: Unstimulated macrophages secrete TNF 516 pg/mL (normally <50). Started with new FBS lot 4141; endotoxin 4.8 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'cell line passage drift' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 0, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [516.0, 4141.0, 4.8, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Incubator CO2 failure", "Cell line passage drift", "Mycoplasma contamination", "Plastics leachables", "Endotoxin-contaminated serum lot activating TLR4", "ELISA cross-reactivity"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Mycoplasma contamination", "ELISA cross-reactivity", "Plastics leachables", "Incubator CO2 failure", "Cell line passage drift"], "domain": "biology_lab", "evidence": "Engineer's handoff: Unstimulated macrophages secrete TNF 1027 pg/mL (normally <50). Started with new FBS lot 4900; endotoxin 1.3 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 4, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [1027.0, 4900.0, 1.3, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["ELISA cross-reactivity", "Cell line passage drift", "Endotoxin-contaminated serum lot activating TLR4", "Mycoplasma contamination", "Plastics leachables", "Incubator CO2 failure"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Plastics leachables", "Incubator CO2 failure", "Mycoplasma contamination", "Cell line passage drift", "ELISA cross-reactivity"], "domain": "biology_lab", "evidence": "Shift log excerpt (cleansed): Unstimulated macrophages secrete TNF 470 pg/mL (normally <50). Started with new FBS lot 4462; endotoxin 4.9 EU/mL by LAL. Old lot cultures normal. Note: shift change at 14:00 badge reader log entry unrelated", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [470.0, 4462.0, 4.9, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Incubator CO2 failure", "Mycoplasma contamination", "ELISA cross-reactivity", "Endotoxin-contaminated serum lot activating TLR4", "Plastics leachables", "Cell line passage drift"], "difficulty": "hard", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Cell line passage drift", "ELISA cross-reactivity", "Mycoplasma contamination", "Incubator CO2 failure", "Plastics leachables"], "domain": "biology_lab", "evidence": "Unstimulated macrophages secrete TNF 961 pg/mL (normally <50). Started with new FBS lot 9979; endotoxin 4.2 EU/mL by LAL. Old lot cultures normal. A colleague suspects 'mycoplasma contamination' because a similar symptom was seen once before (no measurement attached).", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 3, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [961.0, 9979.0, 4.2, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["ELISA cross-reactivity", "Incubator CO2 failure", "Endotoxin-contaminated serum lot activating TLR4", "Plastics leachables", "Mycoplasma contamination", "Cell line passage drift"], "difficulty": "medium", "discriminating_followups": ["Does old FBS lot restore baseline?", "Polymyxin B blocks activation?", "Mycoplasma PCR result?"], "distractors": ["Plastics leachables", "ELISA cross-reactivity", "Cell line passage drift", "Incubator CO2 failure", "Mycoplasma contamination"], "domain": "biology_lab", "evidence": "Engineer's handoff: Unstimulated macrophages secrete TNF 900 pg/mL (normally <50). Started with new FBS lot 4416; endotoxin 4.6 EU/mL by LAL. Old lot cultures normal.", "failure_category": "reagent_material_flaw", "gold_cause": "Endotoxin-contaminated serum lot activating TLR4", "gold_hypothesis_index": 2, "group_id": "fam:BIO-FBS-ENDOTOXIN", "id": "SYN-BIO-FBS-ENDOTOXIN-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["LPS triggers TLR4-mediated TNF secretion at low EU levels."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "cell_culture", "tabular_stats": [900.0, 4416.0, 4.6, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Macrophage activation in control wells", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["On-target knockout causes phenotype", "Clonal selection bottleneck", "Cas9 toxicity", "Mycoplasma", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Antibiotic selection stress"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "Clonal selection bottleneck", "Antibiotic selection stress", "On-target knockout causes phenotype", "Cas9 toxicity"], "domain": "biology_lab", "evidence": "Ticket summary: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 95%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 95.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Clonal selection bottleneck", "Antibiotic selection stress", "Mycoplasma", "Cas9 toxicity", "Off-target cutting of paralog causing phenotype, not on-target knockout", "On-target knockout causes phenotype"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Clonal selection bottleneck", "Antibiotic selection stress", "Mycoplasma", "On-target knockout causes phenotype", "Cas9 toxicity"], "domain": "biology_lab", "evidence": "Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 93%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. duplicate log line suppressed x12 Note: shift change at 14:00", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 93.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Antibiotic selection stress", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Mycoplasma", "Cas9 toxicity", "Clonal selection bottleneck", "On-target knockout causes phenotype"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Cas9 toxicity", "Antibiotic selection stress", "Clonal selection bottleneck", "Mycoplasma", "On-target knockout causes phenotype"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 96%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'on-target knockout causes phenotype' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 1, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 96.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Cas9 toxicity", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Clonal selection bottleneck", "Mycoplasma", "Antibiotic selection stress", "On-target knockout causes phenotype"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "On-target knockout causes phenotype", "Cas9 toxicity", "Clonal selection bottleneck", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Ticket summary: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 89%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 1, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 89.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Antibiotic selection stress", "Off-target cutting of paralog causing phenotype, not on-target knockout", "On-target knockout causes phenotype", "Cas9 toxicity", "Mycoplasma", "Clonal selection bottleneck"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "Clonal selection bottleneck", "Cas9 toxicity", "Antibiotic selection stress", "On-target knockout causes phenotype"], "domain": "biology_lab", "evidence": "Ticket summary: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 93%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. Unrelated: cafeteria chiller alarm acknowledged HMI screensaver restarted", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 1, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 93.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Mycoplasma", "On-target knockout causes phenotype", "Cas9 toxicity", "Clonal selection bottleneck", "Antibiotic selection stress", "Off-target cutting of paralog causing phenotype, not on-target knockout"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["On-target knockout causes phenotype", "Clonal selection bottleneck", "Antibiotic selection stress", "Cas9 toxicity", "Mycoplasma"], "domain": "biology_lab", "evidence": "Shift log excerpt (cleansed): Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 93%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'on-target knockout causes phenotype' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 5, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 93.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Antibiotic selection stress", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Clonal selection bottleneck", "Mycoplasma", "On-target knockout causes phenotype", "Cas9 toxicity"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Antibiotic selection stress", "On-target knockout causes phenotype", "Clonal selection bottleneck", "Cas9 toxicity", "Mycoplasma"], "domain": "biology_lab", "evidence": "Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 94%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 1, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 94.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["On-target knockout causes phenotype", "Clonal selection bottleneck", "Mycoplasma", "Cas9 toxicity", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Antibiotic selection stress"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "Cas9 toxicity", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Shift log excerpt (cleansed): Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 90%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. HMI screensaver restarted Note: shift change at 14:00", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 90.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Mycoplasma", "Antibiotic selection stress", "On-target knockout causes phenotype", "Clonal selection bottleneck", "Cas9 toxicity", "Off-target cutting of paralog causing phenotype, not on-target knockout"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Cas9 toxicity", "Clonal selection bottleneck", "Mycoplasma", "On-target knockout causes phenotype", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Ticket summary: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 93%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'clonal selection bottleneck' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 5, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 93.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Clonal selection bottleneck", "Antibiotic selection stress", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Mycoplasma", "On-target knockout causes phenotype", "Cas9 toxicity"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "Cas9 toxicity", "On-target knockout causes phenotype", "Clonal selection bottleneck", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Ticket summary: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 91%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 2, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 91.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Off-target cutting of paralog causing phenotype, not on-target knockout", "Cas9 toxicity", "Mycoplasma", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Antibiotic selection stress"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Antibiotic selection stress", "Mycoplasma", "Cas9 toxicity", "On-target knockout causes phenotype", "Clonal selection bottleneck"], "domain": "biology_lab", "evidence": "Field report: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 90%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. [INFO] heartbeat ok Firmware banner printed at boot", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 0, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 90.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Cas9 toxicity", "On-target knockout causes phenotype", "Mycoplasma", "Clonal selection bottleneck", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Antibiotic selection stress"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Clonal selection bottleneck", "Antibiotic selection stress", "On-target knockout causes phenotype", "Cas9 toxicity", "Mycoplasma"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 95%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'cas9 toxicity' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 95.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Mycoplasma", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Cas9 toxicity", "Antibiotic selection stress"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Cas9 toxicity", "Mycoplasma", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 95%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 3, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 95.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Cas9 toxicity", "Clonal selection bottleneck", "Antibiotic selection stress", "On-target knockout causes phenotype", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Mycoplasma"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Cas9 toxicity", "On-target knockout causes phenotype", "Clonal selection bottleneck", "Mycoplasma", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 95%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. [INFO] heartbeat ok duplicate log line suppressed x12", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 95.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Mycoplasma", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Clonal selection bottleneck", "Cas9 toxicity", "Antibiotic selection stress", "On-target knockout causes phenotype"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Antibiotic selection stress", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Cas9 toxicity", "Mycoplasma"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 92%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'on-target knockout causes phenotype' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 1, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 92.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Clonal selection bottleneck", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Antibiotic selection stress", "Mycoplasma", "Cas9 toxicity", "On-target knockout causes phenotype"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Clonal selection bottleneck", "Mycoplasma", "Cas9 toxicity", "On-target knockout causes phenotype", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 91%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 1, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 91.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Cas9 toxicity", "On-target knockout causes phenotype", "Antibiotic selection stress", "Clonal selection bottleneck", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Mycoplasma"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Clonal selection bottleneck", "Antibiotic selection stress", "Mycoplasma", "Cas9 toxicity", "On-target knockout causes phenotype"], "domain": "biology_lab", "evidence": "Field report: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 92%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. HMI screensaver restarted badge reader log entry unrelated", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 92.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Antibiotic selection stress", "Clonal selection bottleneck", "Mycoplasma", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Cas9 toxicity", "On-target knockout causes phenotype"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Cas9 toxicity", "On-target knockout causes phenotype", "Antibiotic selection stress", "Clonal selection bottleneck", "Mycoplasma"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 92%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'cas9 toxicity' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 3, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 92.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Clonal selection bottleneck", "Cas9 toxicity", "Mycoplasma", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Antibiotic selection stress", "On-target knockout causes phenotype"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Cas9 toxicity", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 1 mismatches to paralog A2 (identity 92%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 3, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [1.0, 92.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Off-target cutting of paralog causing phenotype, not on-target knockout", "On-target knockout causes phenotype", "Mycoplasma", "Clonal selection bottleneck", "Antibiotic selection stress", "Cas9 toxicity"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Mycoplasma", "Antibiotic selection stress", "Cas9 toxicity", "On-target knockout causes phenotype", "Clonal selection bottleneck"], "domain": "biology_lab", "evidence": "Engineer's handoff: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 90%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. badge reader log entry unrelated [INFO] heartbeat ok", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 0, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 90.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Mycoplasma", "Clonal selection bottleneck", "On-target knockout causes phenotype", "Antibiotic selection stress", "Cas9 toxicity", "Off-target cutting of paralog causing phenotype, not on-target knockout"], "difficulty": "hard", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Clonal selection bottleneck", "Mycoplasma", "On-target knockout causes phenotype", "Cas9 toxicity", "Antibiotic selection stress"], "domain": "biology_lab", "evidence": "Shift log excerpt (cleansed): Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 95%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype. A colleague suspects 'clonal selection bottleneck' because a similar symptom was seen once before (no measurement attached).", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 5, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 95.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Clonal selection bottleneck", "On-target knockout causes phenotype", "Antibiotic selection stress", "Mycoplasma", "Off-target cutting of paralog causing phenotype, not on-target knockout", "Cas9 toxicity"], "difficulty": "medium", "discriminating_followups": ["Amplicon sequencing of paralog site?", "Does paralog cDNA rescue?", "Phenotype with a third guide?"], "distractors": ["Clonal selection bottleneck", "On-target knockout causes phenotype", "Antibiotic selection stress", "Cas9 toxicity", "Mycoplasma"], "domain": "biology_lab", "evidence": "Ticket summary: Team attributes growth defect to knockout of gene A. sgRNA has 2 mismatches to paralog A2 (identity 88%). Rescue with A cDNA does not restore growth; second sgRNA against A shows no phenotype.", "failure_category": "flawed_hypothesis", "gold_cause": "Off-target cutting of paralog causing phenotype, not on-target knockout", "gold_hypothesis_index": 4, "group_id": "fam:BIO-CRISPR-PARALOG", "id": "SYN-BIO-CRISPR-PARALOG-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Failure of cDNA rescue and absence of phenotype with an independent guide argue against on-target cause."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "genome_editing", "tabular_stats": [2.0, 88.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Phenotype attributed to target knockout", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Datum/height reference mismatch between GCPs and processing", "Insufficient image overlap", "GPS multipath", "Rolling shutter", "Camera lens distortion", "Wind drift"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Camera lens distortion", "Rolling shutter", "GPS multipath", "Insufficient image overlap", "Wind drift"], "domain": "sensors", "evidence": "Engineer's handoff: Drone photogrammetry model offset 2.2 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 36 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 0, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-00", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.2, 36.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Rolling shutter", "Camera lens distortion", "Wind drift", "GPS multipath", "Insufficient image overlap", "Datum/height reference mismatch between GCPs and processing"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Camera lens distortion", "Rolling shutter", "GPS multipath", "Insufficient image overlap", "Wind drift"], "domain": "sensors", "evidence": "Field report: Drone photogrammetry model offset 1.4 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 41 m. Firmware banner printed at boot Note: shift change at 14:00", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 5, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-01", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.4, 41.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["GPS multipath", "Insufficient image overlap", "Rolling shutter", "Camera lens distortion", "Datum/height reference mismatch between GCPs and processing", "Wind drift"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Camera lens distortion", "GPS multipath", "Wind drift", "Insufficient image overlap", "Rolling shutter"], "domain": "sensors", "evidence": "Engineer's handoff: Drone photogrammetry model offset 1.9 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 41 m. A colleague suspects 'insufficient image overlap' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 4, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-02", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.9, 41.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Camera lens distortion", "Wind drift", "Datum/height reference mismatch between GCPs and processing", "Insufficient image overlap", "GPS multipath", "Rolling shutter"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Insufficient image overlap", "Camera lens distortion", "Rolling shutter", "Wind drift", "GPS multipath"], "domain": "sensors", "evidence": "Engineer's handoff: Drone photogrammetry model offset 1.2 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 43 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 2, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-03", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.2, 43.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Datum/height reference mismatch between GCPs and processing", "Rolling shutter", "Camera lens distortion", "Insufficient image overlap", "GPS multipath", "Wind drift"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Camera lens distortion", "Wind drift", "Rolling shutter", "Insufficient image overlap", "GPS multipath"], "domain": "sensors", "evidence": "Field report: Drone photogrammetry model offset 2.0 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 22 m. [INFO] heartbeat ok HMI screensaver restarted", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 0, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-04", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.0, 22.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Wind drift", "Datum/height reference mismatch between GCPs and processing", "GPS multipath", "Insufficient image overlap", "Rolling shutter", "Camera lens distortion"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Wind drift", "Insufficient image overlap", "Rolling shutter", "Camera lens distortion", "GPS multipath"], "domain": "sensors", "evidence": "Field report: Drone photogrammetry model offset 1.7 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 27 m. A colleague suspects 'insufficient image overlap' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 1, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-05", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.7, 27.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Wind drift", "Datum/height reference mismatch between GCPs and processing", "GPS multipath", "Camera lens distortion", "Insufficient image overlap", "Rolling shutter"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Camera lens distortion", "Wind drift", "GPS multipath", "Insufficient image overlap", "Rolling shutter"], "domain": "sensors", "evidence": "Engineer's handoff: Drone photogrammetry model offset 1.3 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 22 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 1, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-06", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.3, 22.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Camera lens distortion", "GPS multipath", "Insufficient image overlap", "Wind drift", "Rolling shutter", "Datum/height reference mismatch between GCPs and processing"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["GPS multipath", "Camera lens distortion", "Insufficient image overlap", "Rolling shutter", "Wind drift"], "domain": "sensors", "evidence": "Operator note: Drone photogrammetry model offset 2.7 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 45 m. Firmware banner printed at boot HMI screensaver restarted", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 5, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-07", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.7, 45.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["GPS multipath", "Camera lens distortion", "Insufficient image overlap", "Datum/height reference mismatch between GCPs and processing", "Wind drift", "Rolling shutter"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["GPS multipath", "Insufficient image overlap", "Wind drift", "Rolling shutter", "Camera lens distortion"], "domain": "sensors", "evidence": "Operator note: Drone photogrammetry model offset 2.9 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 42 m. A colleague suspects 'gps multipath' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 3, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-08", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.9, 42.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Insufficient image overlap", "Camera lens distortion", "Wind drift", "GPS multipath", "Datum/height reference mismatch between GCPs and processing", "Rolling shutter"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["GPS multipath", "Rolling shutter", "Insufficient image overlap", "Camera lens distortion", "Wind drift"], "domain": "sensors", "evidence": "Drone photogrammetry model offset 2.3 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 28 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 4, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-09", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.3, 28.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Wind drift", "GPS multipath", "Datum/height reference mismatch between GCPs and processing", "Insufficient image overlap", "Camera lens distortion", "Rolling shutter"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Camera lens distortion", "Wind drift", "GPS multipath", "Rolling shutter", "Insufficient image overlap"], "domain": "sensors", "evidence": "Field report: Drone photogrammetry model offset 1.2 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 22 m. Unrelated: cafeteria chiller alarm acknowledged Note: shift change at 14:00", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 2, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-10", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.2, 22.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Insufficient image overlap", "Wind drift", "Rolling shutter", "Camera lens distortion", "GPS multipath", "Datum/height reference mismatch between GCPs and processing"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Rolling shutter", "Wind drift", "GPS multipath", "Insufficient image overlap", "Camera lens distortion"], "domain": "sensors", "evidence": "Drone photogrammetry model offset 2.7 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 30 m. A colleague suspects 'camera lens distortion' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 5, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-11", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.7, 30.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Insufficient image overlap", "Rolling shutter", "Wind drift", "Camera lens distortion", "Datum/height reference mismatch between GCPs and processing", "GPS multipath"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Wind drift", "Insufficient image overlap", "Rolling shutter", "Camera lens distortion", "GPS multipath"], "domain": "sensors", "evidence": "Ticket summary: Drone photogrammetry model offset 1.4 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 31 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 4, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-12", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.4, 31.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Rolling shutter", "Insufficient image overlap", "Datum/height reference mismatch between GCPs and processing", "Camera lens distortion", "GPS multipath", "Wind drift"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["GPS multipath", "Rolling shutter", "Wind drift", "Insufficient image overlap", "Camera lens distortion"], "domain": "sensors", "evidence": "Drone photogrammetry model offset 2.5 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 25 m. Note: shift change at 14:00 HMI screensaver restarted", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 2, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-13", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.5, 25.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Insufficient image overlap", "GPS multipath", "Camera lens distortion", "Rolling shutter", "Datum/height reference mismatch between GCPs and processing", "Wind drift"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Insufficient image overlap", "GPS multipath", "Camera lens distortion", "Rolling shutter", "Wind drift"], "domain": "sensors", "evidence": "Drone photogrammetry model offset 2.0 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 25 m. A colleague suspects 'insufficient image overlap' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 4, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-14", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.0, 25.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Insufficient image overlap", "Rolling shutter", "Wind drift", "Datum/height reference mismatch between GCPs and processing", "Camera lens distortion", "GPS multipath"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Insufficient image overlap", "Camera lens distortion", "Rolling shutter", "Wind drift", "GPS multipath"], "domain": "sensors", "evidence": "Operator note: Drone photogrammetry model offset 1.8 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 43 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 3, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-15", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.8, 43.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["GPS multipath", "Rolling shutter", "Insufficient image overlap", "Wind drift", "Datum/height reference mismatch between GCPs and processing", "Camera lens distortion"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Insufficient image overlap", "Rolling shutter", "Wind drift", "GPS multipath", "Camera lens distortion"], "domain": "sensors", "evidence": "Drone photogrammetry model offset 1.9 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 33 m. Firmware banner printed at boot Unrelated: cafeteria chiller alarm acknowledged", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 4, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-16", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.9, 33.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Camera lens distortion", "Wind drift", "GPS multipath", "Rolling shutter", "Insufficient image overlap", "Datum/height reference mismatch between GCPs and processing"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["GPS multipath", "Wind drift", "Rolling shutter", "Insufficient image overlap", "Camera lens distortion"], "domain": "sensors", "evidence": "Drone photogrammetry model offset 2.9 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 43 m. A colleague suspects 'gps multipath' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 5, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-17", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.9, 43.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Datum/height reference mismatch between GCPs and processing", "Camera lens distortion", "Insufficient image overlap", "GPS multipath", "Rolling shutter", "Wind drift"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Insufficient image overlap", "Camera lens distortion", "Rolling shutter", "GPS multipath", "Wind drift"], "domain": "sensors", "evidence": "Ticket summary: Drone photogrammetry model offset 2.5 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 31 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 0, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-18", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.5, 31.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Camera lens distortion", "Insufficient image overlap", "GPS multipath", "Datum/height reference mismatch between GCPs and processing", "Wind drift", "Rolling shutter"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Insufficient image overlap", "GPS multipath", "Camera lens distortion", "Rolling shutter", "Wind drift"], "domain": "sensors", "evidence": "Operator note: Drone photogrammetry model offset 2.7 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 49 m. Note: shift change at 14:00 badge reader log entry unrelated", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 3, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-19", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.7, 49.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["GPS multipath", "Rolling shutter", "Insufficient image overlap", "Datum/height reference mismatch between GCPs and processing", "Camera lens distortion", "Wind drift"], "difficulty": "hard", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Rolling shutter", "Insufficient image overlap", "GPS multipath", "Wind drift", "Camera lens distortion"], "domain": "sensors", "evidence": "Shift log excerpt (cleansed): Drone photogrammetry model offset 1.6 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 41 m. A colleague suspects 'camera lens distortion' because a similar symptom was seen once before (no measurement attached).", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 3, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-20", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [1.6, 41.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["GPS multipath", "Camera lens distortion", "Wind drift", "Datum/height reference mismatch between GCPs and processing", "Insufficient image overlap", "Rolling shutter"], "difficulty": "medium", "discriminating_followups": ["Are horizontal and vertical datums consistent?", "Does geoid correction remove vertical offset?", "Is offset uniform across site?"], "distractors": ["Wind drift", "Camera lens distortion", "Rolling shutter", "GPS multipath", "Insufficient image overlap"], "domain": "sensors", "evidence": "Operator note: Drone photogrammetry model offset 2.6 m from GCPs uniformly. GCPs surveyed in local datum; processing assumed WGS84 ellipsoid heights; vertical offset 48 m.", "failure_category": "protocol_human_loophole", "gold_cause": "Datum/height reference mismatch between GCPs and processing", "gold_hypothesis_index": 3, "group_id": "fam:SEN-DATUM", "id": "SYN-SEN-DATUM-21", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Uniform offsets of known datum-shift magnitude indicate reference frame mismatch."], "source_file": null, "source_type": "synthetic_template", "split": "validation", "subdomain": "geospatial", "tabular_stats": [2.6, 48.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Survey offset of several meters", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": ["Far-field distance 2D^2/lambda is 0.53 m", "The team applied Friis at 0.42 m in a short chamber", "Aperture phase curvature was 38 deg", "Apparent Friis gain read 2.1 dBi", "The same reduction at 1.10 m and a compact range both give ~6.5 dBi", "Substrate-lot and mismatch stories do not produce this range curve"], "competing_hypotheses": ["Horn standard is 4 dB off", "The measurement hypothesis \u2014 that received power follows inverse-square Friis at 0.42 m \u2014 is invalid in the radiative near field.", "The new substrate lot lowered gain", "Uncorrected cable loss", "Chamber multipath"], "difficulty": "hard", "discriminating_followups": ["Check: Apparent gain versus r is not constant below 0.8 m", "Check: 2D^2/lambda = 0.53 m > 0.42 m", "Check: 38 deg aperture phase curvature", "Check: 1.10 m and compact-range gains agree"], "distractors": ["The new substrate lot lowered gain", "Uncorrected cable loss", "Chamber multipath", "Horn standard is 4 dB off"], "domain": "sensors", "evidence": "Measure 2.45 GHz patch-antenna gain in an anechoic corner to +/-0.6 dB using a 10 dBi standard horn. The team used Friis / inverse-square at 0.42 m separation and reported 2.1 dBi instead of the expected 6.5 dBi. Step 1: Place the AUT and horn 0.42 m apart, boresight aligned. AUT aperture D = 0.18 m; far-field criterion 2D^2/lambda is 0.53 m. They used 0.42 m 'because the chamber is small'. A range equation spreadsheet assumed far-field Friis. Step 2: Record S21 and convert to gain with Friis. Cable and mismatch corrections are applied; they are not the 4 dB miss. S21 versus distance from 0.25-1.20 m was later collected on a rail. Step 3: Compare to a 1.10 m far-field point and a compact-range borrow. At 1.10 m the same Friis reduction gives 6.4 dBi; a compact range gives 6.6 dBi. [gain_vs_range] Apparent gain is a strong function of r below 0.8 m. [gain_vs_range] It asymptotes near 6.4 dBi once r exceeds ~1.5 x 2D^2/lambda. [phase] At 0.42 m the AUT aperture phase varies 38 deg peak-to-peak. [phase] At 1.10 m the variation is 9 deg. [phase] A spherical-wave / 1/r^2 assumption is invalid with that curvature. [rf_integrity] Mismatch and cable terms are already in the reduction. [rf_integrity] Chamber reflectivity cannot explain a smooth 4 dB range dependence. 2D^2/lambda at 2.45 GHz for D = 0.18 m is 0.53 m; they measured at 0.42 m. Friis assumes plane-wave / far-field incidence. The chamber length limited the first setup. A student argued the patch 'just has low gain' because of a new substrate lot.", "failure_category": "flawed_hypothesis", "gold_cause": "The measurement hypothesis \u2014 that received power follows inverse-square Friis at 0.42 m \u2014 is invalid in the radiative near field.", "gold_hypothesis_index": 1, "group_id": "yaml:FH-021", "id": "EXP-FH-021", "intervention": "Repeat S21 on a rail from 0.3-1.2 m and reduce gain only for r >= 1.0 m (or run a near-field scan and transform). If gain asymptotes to 6.5 dBi without changing the AUT, the far-field hypothesis is rejected. Do not scrap the substrate lot first.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/flawed_hypothesis/FH-021_near_field_inverse_square.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "rf_metrology", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Antenna Gain Error from Applying Inverse-Square in the Near Field", "why_alternatives_fail": "{\"The new substrate lot lowered gain\": \"The same AUT measures 6.4-6.6 dBi at a legal range.\", \"Uncorrected cable loss\": \"Cable loss is applied and would not create a smooth range-dependent gain.\", \"Chamber multipath\": \"Reflectivity is -38 dB and the curve is monotonic with r, not a ripple.\", \"Horn standard is 4 dB off\": \"The horn plus compact-range comparison agrees at far field.\"}"}
{"category_label": "flawed_hypothesis", "causal_chain": ["The team hypothesized that the new electrolyte causes catastrophic cycle fade", "Cycling to cycle 80 was actually healthy for both electrolytes", "Cells then sat 11 weeks at 100% SOC and 40 C", "Calendar aging consumed lithium inventory in both chemistries", "Post-gap capacity sat near 45% before further cycling", "A 30% SOC / 10 C control barely faded, isolating storage, not the molecule"], "competing_hypotheses": ["The scientific hypothesis \u2014 that post-gap fade is a cycle-life failure of the new electrolyte \u2014 is wrong.", "Cycler current-sensor drift", "Formation was incomplete", "A wet glovebox poisoned only the new lot", "The new electrolyte etches the NMC"], "difficulty": "hard", "discriminating_followups": ["Check: Baseline and new electrolyte drop together", "Check: dQ/dV indicates LLI, not a unique cathode collapse", "Check: Withheld: 77 days, 40 C, 100% SOC", "Check: Fridge control at 30% SOC loses 2%"], "distractors": ["The new electrolyte etches the NMC", "Cycler current-sensor drift", "A wet glovebox poisoned only the new lot", "Formation was incomplete"], "domain": "electrical", "evidence": "Show that a new electrolyte raises 25 C cycle life to 80% capacity at >600 cycles versus a 400-cycle baseline. After a summer storage gap, both chemistries sat at 42-45% capacity before the first post-gap cycle. Step 1: Form cells and begin C/3 cycling at 25 C. Both new and baseline electrolytes reached cycle 80 near 96% capacity. The cycler lab shut for 11 weeks. Cells were left in a 40 C formation oven at 100% SOC 'to keep them dry'. Step 2: Resume C/3 cycling after the shutdown. First discharge after storage is 43-46% for both electrolytes. The team attributed the drop to the new electrolyte attacking the cathode. Step 3: Run dQ/dV and electrode harvest on one cell. Loss of lithium inventory dominates; transition-metal dissolution is minor. [capacity] Both chemistries lose ~50 points during the gap, then fade slowly again. [capacity] The drop is not unique to the new electrolyte. [dqdv] Graphite staging peaks lose area (LLI) with little peak shift that would indicate a bulk cathode crystal change. [dqdv] A harvested anode shows a thick SEI; cathode lattice parameters are in family. [storage_record] 77 days at 40 C and 100% SOC is a known calendar-aging extreme. Cells sat 11 weeks at 100% SOC in a 40 C oven. Calendar aging of NMC/graphite at high SOC is often larger than modest cycle counts. The new electrolyte's selling point was cycle life, so the team framed every fade as a cycle-life failure. A control that stayed at 30% SOC in a 10 C fridge over the same gap lost only 2%.", "failure_category": "flawed_hypothesis", "gold_cause": "The scientific hypothesis \u2014 that post-gap fade is a cycle-life failure of the new electrolyte \u2014 is wrong.", "gold_hypothesis_index": 0, "group_id": "yaml:FH-016", "id": "EXP-FH-016", "intervention": "Build matched cells: cycle continuously at 25 C, versus store 11 weeks at 100% SOC / 40 C versus 30% SOC / 10 C, for both electrolytes. If storage condition predicts fade and electrolyte identity does not, the cycle-life hypothesis is rejected.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/incomplete/FH-016_incomplete_calendar_aging.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "battery_aging", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Incomplete Aging Study: Calendar Fade Misread as Cycle-Life Failure", "why_alternatives_fail": "{\"The new electrolyte etches the NMC\": \"The baseline bottle fades the same amount over the same gap.\", \"Cycler current-sensor drift\": \"dQ/dV shape and harvested SEI show real LLI.\", \"A wet glovebox poisoned only the new lot\": \"Both lots sit at 45% after shared storage.\", \"Formation was incomplete\": \"Cycle 80 was 96% for both before the gap.\"}"}
{"category_label": "reagent_material_flaw", "causal_chain": ["A nearby rail grinder threw ferrous swarf into the shared aisle", "Swarf entered the vented cover of the joint-4 magnetic encoder", "Embedded particles locally distorted the ring-magnet field", "The encoder reported extra mid-travel counts that the controller treated as true angle", "Tool-flange yaw drifted through the shift and pouch placement failed", "An independent laser tracker showed the arm itself was not moving as reported"], "competing_hypotheses": ["Laser tracker nest moved", "Ferrous grinding swarf contaminated the magnetic encoder ring on joint 4.", "Harmonic-drive backlash increased", "Bus-voltage sag caused missing encoder clocks", "Teach-pendant firmware corrupted joint 4 kinematics"], "difficulty": "hard", "discriminating_followups": ["Check: Laser tracker says the arm is smooth while the encoder is not", "Check: Error is joint-4 specific and peaks at mid-travel", "Check: Gearbox backlash and temperature are normal", "Check: Error grows during a shift that included nearby grinding (withheld from the prompt)"], "distractors": ["Harmonic-drive backlash increased", "Teach-pendant firmware corrupted joint 4 kinematics", "Laser tracker nest moved", "Bus-voltage sag caused missing encoder clocks"], "domain": "robotics", "evidence": "Keep a six-axis pick cell's tool-flange pose within 0.15 mm of the taught waypoint after a 4-hour production shift. Overnight, joint 4 began reporting a slow 0.8 deg drift and the cell started dropping pouches at the place nest. Step 1: Home the robot and run the standard 12-point circularity check. Absolute magnetic encoder on joint 4, 19-bit, ring magnet bonded to the output shaft. Morning circularity was 0.09 mm; after four hours it opened to 0.41 mm, biased in yaw. Step 2: Compare encoder counts to an independent laser tracker at the flange. Laser tracker is on a calibrated nest 1.8 m from the cell. True flange motion is smooth; the encoder unwraps extra counts near mid-travel. Step 3: Inspect joint 4 after the shift. No gearbox backlash change on the lock-to-lock test; grease looks normal. A technician wiped a gray film off the encoder cover vent but did not bag the residue. [pose_error] Error grows through the shift rather than appearing at power-up. [pose_error] The growth is almost entirely joint-4 yaw, not a base-frame shift. [encoder_vs_tracker] Disagreement peaks near the mid-travel zone, not at the hard stops. [encoder_vs_tracker] Repeating the same move without power cycling reproduces the hump. [gearbox_check] Backlash, temperature, and bus voltage stay in-family. A portable rail grinder was used on the neighboring cell for two hours that morning; sparks were visible over the shared aisle. Joint 4's encoder is a magnetic ring with a vented cover, not an optical disk. The teach pendant firmware was updated last month and has been stable on the other five joints. Payload mass and TCP calibration were not changed. The laser tracker battery was swapped at lunch; the afternoon survey still matches the morning tracker scale.", "failure_category": "reagent_material_flaw", "gold_cause": "Ferrous grinding swarf contaminated the magnetic encoder ring on joint 4.", "gold_hypothesis_index": 1, "group_id": "yaml:RF-017", "id": "EXP-RF-017", "intervention": "Power down, remove the encoder cover, clean the ring and read-head with a non-residue solvent, and inspect for embedded swarf. Recalibrate joint 4 and rerun the circularity check. If the mid-travel encoder-minus-tracker hump collapses below 0.05 deg, the contamination diagnosis is confirmed. Add a sealed encoder or a magnetic shield when grinding is scheduled nearby.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/incomplete/RF-017_incomplete_encoder_swarf.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "joint_sensing", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Incomplete Robot Cell: Magnetic Encoder Drift After Nearby Grinding", "why_alternatives_fail": "{\"Harmonic-drive backlash increased\": \"Lock-to-lock backlash is 1.8 arcmin, inside spec, and the error is a mid-travel hump rather than endpoint slop.\", \"Teach-pendant firmware corrupted joint 4 kinematics\": \"The other five joints and the morning check were fine; firmware does not grow a mid-travel magnetic hump over four hours.\", \"Laser tracker nest moved\": \"The nest is fixed and afternoon scale still matches morning; the disagreement is encoder-specific.\", \"Bus-voltage sag caused missing encoder clocks\": \"Bus voltage is 48.1 V and a missing-clock fault would step, not form a repeatable mid-travel curve.\"}"}
{"category_label": "instrumentation_artifact", "causal_chain": ["Maintenance disturbed the compressor pressure sensor connector", "The installed pressure channel developed a low bias", "The RUL model consumed the biased channel as if it were physical engine behavior", "Low compressor pressure residual mimicked compressor efficiency degradation", "Other channels failed to corroborate true degradation", "The model predicted a false remaining-useful-life collapse"], "competing_hypotheses": ["RUL model architecture is invalid", "Fuel control problem", "The installed compressor discharge pressure sensor was biased low after its harness was disturbed during maintenance.", "Actual compressor efficiency degradation"], "difficulty": "hard", "discriminating_followups": ["Check: Installed pressure sensor reads 7.5% low against a reference", "Check: Temperature and fuel flow do not corroborate engine degradation", "Check: Fault appears immediately after connector disturbance", "Check: Borescope inspection found no physical compressor damage"], "distractors": ["Actual compressor efficiency degradation", "Fuel control problem", "RUL model architecture is invalid"], "domain": "mechanical", "evidence": "Diagnose why a turbofan prognostics model predicts a rapid remaining-useful-life drop during a ground test. The engine should show stable health indicators after a routine borescope inspection. Step 1: Run a standard ground-test profile after maintenance. Sensors include fan speed, compressor discharge pressure, exhaust gas temperature, and fuel flow. Step 2: Feed normalized sensor channels into the RUL estimator. The estimator was trained on C-MAPSS-like degradation trajectories. Step 3: Cross-check suspect channels against redundant gauges. A portable calibrated pressure transducer is installed after the model flags rapid degradation. [prognostics_output] RUL collapses mainly because compressor pressure appears low. [prognostics_output] Temperature and fuel-flow channels do not show a matching degradation pattern. [calibration_check] Installed sensor reads about 7.5% low. [calibration_check] Portable reference matches expected pressure. [maintenance_notes] Pressure sensor harness was disconnected during borescope access. [maintenance_notes] Connector shell shows a bent shield clamp and slight moisture residue. [maintenance_notes] Borescope found no compressor blade damage. The RUL model assumes sensor channels are calibrated and comparable to training data. True compressor degradation should also shift exhaust gas temperature and fuel flow. The warning appeared immediately after maintenance, not gradually over flight cycles. The portable pressure transducer was calibrated last month.", "failure_category": "instrumentation_artifact", "gold_cause": "The installed compressor discharge pressure sensor was biased low after its harness was disturbed during maintenance.", "gold_hypothesis_index": 2, "group_id": "yaml:IN-012", "id": "EXP-IN-012", "intervention": "Repair or reseat the pressure-sensor harness, recalibrate the channel, rerun the ground test, and compare against the portable reference. The RUL estimate should return to the stable range if the root cause is the pressure bias.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/instrumentation/IN-012_turbofan_sensor_bias_false_degradation.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "engine_prognostics", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Turbofan Health Monitor Flags False Degradation from Compressor Sensor Bias", "why_alternatives_fail": "{\"Actual compressor efficiency degradation\": \"Reference pressure and borescope inspection do not support physical degradation.\", \"Fuel control problem\": \"Fuel-flow channel is stable.\", \"RUL model architecture is invalid\": \"The immediate issue is a biased input channel, not the model structure.\"}"}
{"category_label": "instrumentation_artifact", "causal_chain": ["A high-concentration standard was injected before blanks and study samples", "The sticky analyte remained in the needle or injection path", "The 50% methanol wash did not remove all residual analyte", "Residual analyte eluted in subsequent blank and sample injections", "Blank signal decayed with injection order and exceeded LLOQ", "Low-level study positives were artifacts of carryover"], "competing_hypotheses": ["True low-level biomarker is present in blanks", "Matrix effect suppresses internal standard", "The LC-MS/MS method has autosampler/needle carryover after high-concentration standards.", "Calibration curve is nonlinear"], "difficulty": "medium", "discriminating_followups": ["Check: Blank signal decays after high standard", "Check: Strong wash eliminates blank signal", "Check: False positives cluster after high injections", "Check: Retention time and ion ratio match residual standard"], "distractors": ["True low-level biomarker is present in blanks", "Matrix effect suppresses internal standard", "Calibration curve is nonlinear"], "domain": "biology_lab", "evidence": "Quantify a low-abundance drug metabolite in plasma samples by LC-MS/MS. Blank samples should be below the lower limit of quantification. Step 1: Inject calibration standards from high to low concentration. Highest standard is 5000 ng/mL, followed by blanks and study samples. Step 2: Run the study batch with the validated gradient. Autosampler needle wash uses 50% methanol. Step 3: Review blanks and reinject after stronger needle wash. A 90% acetonitrile/isopropanol wash is tested after unexpected positives. [batch_sequence] Blank peak area decays after the high standard. [batch_sequence] Low-positive samples are near the carryover tail. [wash_test] Strong organic wash pushes blank signal below LLOQ. The suspected biomarker is sticky and hydrophobic. The high standard was injected before blanks rather than after blanks. Retention time and product-ion ratio in blanks match the high standard. Patient samples with positives cluster after high-concentration injections.", "failure_category": "instrumentation_artifact", "gold_cause": "The LC-MS/MS method has autosampler/needle carryover after high-concentration standards.", "gold_hypothesis_index": 2, "group_id": "yaml:IN-013", "id": "EXP-IN-013", "intervention": "Reorder the batch with blanks before high standards, use a strong organic needle wash, and require blank-after-high-standard carryover below 20% of LLOQ. Reinject affected samples after clean blanks to confirm true positives.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/instrumentation/IN-013_lcms_carryover_false_biomarker.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "mass_spectrometry", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "LC-MS Carryover Creates False Low-Level Biomarker Positives", "why_alternatives_fail": "{\"True low-level biomarker is present in blanks\": \"Blank signal depends on injection order after high standards.\", \"Matrix effect suppresses internal standard\": \"The artifact appears in solvent blanks as well as samples.\", \"Calibration curve is nonlinear\": \"Changing needle wash removes the blank peaks without changing calibration model.\"}"}
{"category_label": "instrumentation_artifact", "causal_chain": ["Warm water and delayed maintenance allowed algae to grow on the optical window", "Biofilm and trapped silt increased light scattering inside the probe path", "The probe interpreted window fouling as high suspended sediment", "Field NTU rose while manual grab samples remained near baseline", "Cleaning the window restored normal readings"], "competing_hypotheses": ["Construction runoff event", "Benchtop turbidimeter failed", "Natural storm-driven sediment pulse", "The turbidity spike is an optical measurement artifact caused by biofouling and silt on the probe window."], "difficulty": "easy", "discriminating_followups": ["Check: Manual grab samples are normal", "Check: No hydrologic trigger is present", "Check: Probe window is visibly fouled", "Check: Cleaning immediately restores baseline"], "distractors": ["Construction runoff event", "Benchtop turbidimeter failed", "Natural storm-driven sediment pulse"], "domain": "sensors", "evidence": "Monitor river turbidity downstream of a construction site. Regulatory alerts trigger if turbidity exceeds baseline by more than 50 NTU for over 30 minutes. Step 1: Deploy optical turbidity probe on a fixed river platform. Probe reports 5-minute NTU averages and is cleaned every two weeks. Step 2: Investigate a sudden turbidity alert. The alert occurs after 12 days without rain. Step 3: Collect grab samples and inspect the probe window. Manual samples are measured on a benchtop turbidimeter. [field_probe] Probe reports a sharp turbidity increase without rainfall or flow increase. [grab_samples] Manual samples remain near baseline. [inspection] Optical window is covered by algae film and trapped silt. [inspection] After cleaning, field probe immediately reads 21 NTU. Water temperature increased during the week before the alert. No rainfall, excavation, or upstream flow change is recorded. The cleaning interval was extended because staff were unavailable. A second upstream station did not show a turbidity spike.", "failure_category": "instrumentation_artifact", "gold_cause": "The turbidity spike is an optical measurement artifact caused by biofouling and silt on the probe window.", "gold_hypothesis_index": 3, "group_id": "yaml:IN-014", "id": "EXP-IN-014", "intervention": "Clean the probe, redeploy with an upstream duplicate or wiper-equipped sensor, and compare field readings to grab samples after the same weather conditions.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/instrumentation/IN-014_turbidity_sensor_biofouling.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "water_quality_monitoring", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "River Turbidity Spike from Biofouled Optical Sensor Window", "why_alternatives_fail": "{\"Construction runoff event\": \"No rain, flow change, or upstream station spike occurred.\", \"Benchtop turbidimeter failed\": \"Multiple grab samples agree with the upstream station.\", \"Natural storm-driven sediment pulse\": \"There was no rainfall or discharge increase.\"}"}
{"category_label": "protocol_human_loophole", "causal_chain": ["The night shift packed twice the SOP load density", "The soak clock started on the air thermocouple, not a part thermocouple", "Center parts peaked near 801 C instead of 845 C", "Cores did not fully austenitize", "Oil quench produced mixed transformation products", "Center cores tempered out at 43-46 HRC while edges met 50-53 HRC"], "competing_hypotheses": ["Severe decarburization", "Temper overshot and softened everything", "The 4340 heat is off-chemistry", "The operator double-stacked 48 blocks and started the 40-minute soak when the air thermocouple hit set-point, not when the heaviest part did.", "Quench oil was too hot or watery"], "difficulty": "medium", "discriminating_followups": ["Check: Hardness depends on basket position, not heat number", "Check: Center cores show mixed transformation, not over-temper", "Check: Same heat passed in a 24-piece load", "Check: Instrumented dummy in the dense pack peaks at 801 C"], "distractors": ["Quench oil was too hot or watery", "The 4340 heat is off-chemistry", "Temper overshot and softened everything", "Severe decarburization"], "domain": "manufacturing_industrial", "evidence": "Through-harden 4340 coupon blocks to 50-53 HRC with a 845 C soak and oil quench. Surface and mid-thickness hardness should agree within 2 HRC. Step 1: Load 48 blocks into a batch furnace basket. SOP allows 24 blocks with 15 mm spacing. The night shift double-stacked to finish the order. The furnace air thermocouple reached 845 C in 38 minutes; no part-core thermocouple was used. Step 2: Soak 40 minutes from the air-TC set-point, then oil quench. SOP soak is 40 minutes after the heaviest part is within 10 C of set-point. The operator started the soak clock when the air TC first hit 845 C. Step 3: Temper 200 C / 2 h and hardness-map six blocks from edge, middle, and center of the basket. Edge blocks pass; center blocks are 42-46 HRC in the core. [hardness] Basket-edge parts are through-hard. [hardness] Basket-center cores are 6-8 HRC soft. [furnace] Oil temperature is in family. [furnace] A later instrumented dummy in the same packing peaks at 801 C when the air TC is already at 845 C. [microstructure] Mixed martensite plus fine pearlite/bainite in the core. [microstructure] Edge coupons are nearly full martensite. [microstructure] No decarburization deeper than 0.04 mm. The SOP sketch shows a single layer with 15 mm gaps. The same 4340 heat through-hardened last week in a 24-piece load. Quench oil was analyzed last month and is within water and soot limits. A new temper furnace was commissioned, so some blamed temper color.", "failure_category": "protocol_human_loophole", "gold_cause": "The operator double-stacked 48 blocks and started the 40-minute soak when the air thermocouple hit set-point, not when the heaviest part did.", "gold_hypothesis_index": 3, "group_id": "yaml:PL-011", "id": "EXP-PL-011", "intervention": "Reload 24 blocks with SOP spacing, embed a core thermocouple in a dummy, start the soak only after the dummy is within 10 C of 845 C, and map hardness again. If center cores rise to 50-53 HRC, the packing/soak-clock diagnosis is confirmed.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/protocol_loophole/PL-011_heat_treat_dense_load.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "heat_treatment", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Soft Cores from a Heat-Treat Load Packed Too Densely", "why_alternatives_fail": "{\"Quench oil was too hot or watery\": \"Oil is 62 C and in spec; edge parts through-harden in the same quench.\", \"The 4340 heat is off-chemistry\": \"Last week's 24-piece load from the same heat passed.\", \"Temper overshot and softened everything\": \"Edge parts are 51-52 HRC after the same temper.\", \"Severe decarburization\": \"Decarb is 0.04 mm; the soft reading is mid-plane.\"}"}
{"category_label": "reagent_material_flaw", "causal_chain": ["Tool EM-6-F3 was reused from a previous run without resetting or checking tool life", "The end mill accumulated 18.4 cutting hours, exceeding the recommended 8 hour life", "Flank wear reached 0.34 mm and one flute developed small edge chipping", "The dull tool required higher cutting force and generated more vibration", "Higher force caused cutter deflection and rubbing during slot finishing", "Slots became undersized while surface roughness and burr height increased", "Late-batch parts failed dimensional and surface-finish inspection"], "competing_hypotheses": ["The CNC program used the wrong cutter compensation", "Coolant concentration was too low", "Fixture shifted during the run", "The reused 6 mm carbide end mill was worn beyond its tool-life limit.", "Material stock hardness was out of specification"], "difficulty": "medium", "discriminating_followups": ["Check: Failure is gradual across part number, consistent with wear rather than setup error", "Check: The same material lot, fixture, program, coolant, and work offset are shared by passing and failing parts", "Check: Spindle load and vibration increase as dimensions drift", "Check: Flank wear exceeds the 0.20 mm replacement threshold"], "distractors": ["Material stock hardness was out of specification", "Fixture shifted during the run", "Coolant concentration was too low", "The CNC program used the wrong cutter compensation"], "domain": "manufacturing_industrial", "evidence": "Machine a batch of aluminum 6061-T6 sensor brackets with a 6.00 +/- 0.03 mm mounting slot, surface roughness Ra below 1.6 um, and no visible burrs. The first 80 parts passed inspection, but parts from the second half of the batch failed dimensional and surface-finish acceptance. Step 1: Load certified 6061-T6 plate stock into the Haas VF-2 CNC mill. Material certificate lists yield strength 276 MPa and hardness inside expected range. The same stock lot was used for passing and failing parts. Step 2: Install 6 mm three-flute carbide end mill for slot finishing. Tool ID EM-6-F3 was reused from the previous production run. Operator did not reset the tool-life counter because the tool still looked visually acceptable. Step 3: Run the validated slot-finishing toolpath. Spindle 12000 rpm, feed 950 mm/min, 0.4 mm radial engagement, flood coolant. Program, fixture, coolant concentration, and work offset match the validated setup sheet. Step 4: Inspect every twentieth part on the CMM and surface profilometer. Slot width, burr height, and Ra are measured after deburring is intentionally skipped for process validation. Measurements drift gradually after part 90 rather than failing abruptly. Step 5: Inspect the end mill under a toolmaker microscope after the failed batch. Measure flank wear land and check cutting-edge chipping. Flank wear measures 0.34 mm with small edge chipping on flute 2. [dimensional_trend] Slot width drifts below the lower tolerance after approximately part 100. [dimensional_trend] The drift is gradual and monotonic rather than random. [surface_finish] Ra exceeds the 1.6 um limit beginning near part 100. [surface_finish] Burr height increases with surface roughness. [machine_load] Spindle load and vibration rise as the batch progresses. [machine_load] Coolant flow and spindle speed remain stable. [tool_inspection] The end mill is more than twice its recommended tool-life. [tool_inspection] Measured flank wear exceeds the replacement threshold. The same plate stock lot produced both passing early parts and failing late parts. Coolant concentration was 7.8%, inside the 7-9% process window. The fixture was probed before and after the run; work offset shifted by less than 0.005 mm. The tool had been used for an earlier abrasive prototype run but the operator's note was not copied into the production traveler. The CNC program hash matches the previously validated process plan.", "failure_category": "reagent_material_flaw", "gold_cause": "The reused 6 mm carbide end mill was worn beyond its tool-life limit.", "gold_hypothesis_index": 3, "group_id": "yaml:RF-009", "id": "EXP-RF-009", "intervention": "Replace EM-6-F3 with a new end mill, rerun the same toolpath on the same stock and fixture for a short confirmation batch, and inspect slot width, Ra, burr height, spindle load, and vibration. If new-tool parts return to 6.00 +/- 0.03 mm with Ra below 1.6 um and lower spindle load, the diagnosis is confirmed. Add tool-life lockout or mandatory microscope inspection before production runs.", "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": "test_cases/reagent_flaw/RF-009_cnc_end_mill_wear.yaml", "source_type": "expert_authored_synthetic", "split": "validation", "subdomain": "cnc_milling", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "CNC End Mill Wear Drives Dimensional Drift and Burr Formation", "why_alternatives_fail": "{\"Material stock hardness was out of specification\": \"The same certified stock lot produced passing early parts and failing late parts.\", \"Fixture shifted during the run\": \"Work offset changed less than 0.005 mm, far below the observed slot drift.\", \"Coolant concentration was too low\": \"Coolant stayed inside the validated 7-9% window and would not explain monotonic tool-load increase alone.\", \"The CNC program used the wrong cutter compensation\": \"The program hash matches the validated plan, and early parts were in tolerance.\"}"}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Process-temperature probe lag under-reporting while cooling is actually adequate", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Power overload from high torque at high RPM rather than cooling deficit", "Process physics consistent with a different AI4I failure mode than the label", "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "Tool wear failure: elevated torque from dull cutter misread as thermal trip"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Power overload from high torque at high RPM rather than cooling deficit", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '298.2', 'Process temperature [K]': '308.3', 'Rotational speed [rpm]': '1345', 'Torque [Nm]': '48.8', 'Tool wear [min]': '74', 'Failure Type': 'Heat Dissipation Failure', 'Machine failure': '1'}. Failure type field=Heat Dissipation Failure.", "failure_category": "reagent_material_flaw", "gold_cause": "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "gold_hypothesis_index": 4, "group_id": "harvest:HV-SYN-0009", "id": "HV-SYN-0009-X1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [298.2, 308.3, 1345.0, 48.8, 74.0, 1.0, 345.8833333333334, 1.0, 1345.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 9", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Heat dissipation path obstructed / cooling insufficient at low rotational speed", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Process physics consistent with a different AI4I failure mode than the label", "Power overload from high torque at high RPM rather than cooling deficit", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Tool wear failure: elevated torque from dull cutter misread as thermal trip"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Power overload from high torque at high RPM rather than cooling deficit", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Industrial sensor readings extract (AI4I-style). Key readings: {'Air temperature [K]': '302.0', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1294', 'Torque [Nm]': '52.6', 'Tool wear [min]': '26', 'Failure Type': 'Heat Dissipation Failure', 'Machine failure': '1'}. Labeled fault class=Heat Dissipation Failure.", "failure_category": "reagent_material_flaw", "gold_cause": "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "gold_hypothesis_index": 0, "group_id": "harvest:HV-SYN-0021", "id": "HV-SYN-0021-X1-P2", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [302.0, 311.8, 1294.0, 52.6, 26.0, 1.0, 331.23333333333335, 1.0, 1294.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 21", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: M", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.0', 'Process temperature [K]': '309.7', 'Rotational speed [rpm]': '1386', 'Torque [Nm]': '62.7', 'Tool wear [min]': '142', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0019", "id": "HV-AI4I-0019-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4354.0, 302.0, 309.7, 1386.0, 62.7, 142.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 546.5333333333333, 0.0, 4354.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 19", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Heat dissipation path obstructed at low rotational speed", "Tool wear beyond process window elevating cutting torque", "Overstrain recipe outside qualified envelope without power-limit hit", "Power/torque path indicates overload trip condition", "Current-transformer phase error inventing apparent power trip"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Heat dissipation path obstructed at low rotational speed", "Tool wear beyond process window elevating cutting torque", "Overstrain recipe outside qualified envelope without power-limit hit", "Current-transformer phase error inventing apparent power trip", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '299.5', 'Process temperature [K]': '308.6', 'Rotational speed [rpm]': '1541', 'Torque [Nm]': '61.8', 'Tool wear [min]': '13', 'Failure Type': 'Power Failure', 'Machine failure': '1'}. Failure type field=Power Failure.", "failure_category": "instrumentation_artifact", "gold_cause": "Power/torque path indicates overload trip condition", "gold_hypothesis_index": 4, "group_id": "harvest:HV-SYN-0015", "id": "HV-SYN-0015-X1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [299.5, 308.6, 1541.0, 61.8, 13.0, 1.0, 370.81666666666666, 1.0, 1541.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 15", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only", "Machine failure label: M"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '303.7', 'Process temperature [K]': '312', 'Rotational speed [rpm]': '2663', 'Torque [Nm]': '11.4', 'Tool wear [min]': '71', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 5, "group_id": "harvest:HV-AI4I-0062", "id": "HV-AI4I-0062", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4763.0, 303.7, 312.0, 2663.0, 11.4, 71.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 677.1750000000001, 0.0, 4763.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 62", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Process physics consistent with a different AI4I failure mode than the label", "Power overload from high torque at high RPM rather than cooling deficit", "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Power overload from high torque at high RPM rather than cooling deficit", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.0', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1294', 'Torque [Nm]': '52.6', 'Tool wear [min]': '26', 'Failure Type': 'Heat Dissipation Failure', 'Machine failure': '1'}. Failure type field=Heat Dissipation Failure.", "failure_category": "reagent_material_flaw", "gold_cause": "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "gold_hypothesis_index": 4, "group_id": "harvest:HV-SYN-0021", "id": "HV-SYN-0021-X1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [302.0, 311.8, 1294.0, 52.6, 26.0, 1.0, 331.23333333333335, 1.0, 1294.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 21", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Ferrous grinding swarf contaminated the magnetic encoder ring on joint 4. Local field distortion added a mid-travel count error that looked like kinematic drift. The mechanical gearbox is healthy; the withheld clue is the neighboring rail-grinder work over an unsealed magnetic encoder.\n", "Harmonic-drive backlash increased", "Laser tracker nest moved", "Teach-pendant firmware corrupted joint 4 kinematics", "encoder contamination", "Bus-voltage sag caused missing encoder clocks"], "difficulty": "hard", "discriminating_followups": ["What evidence exists regarding 'tracker'?", "What evidence exists regarding 'mid-travel'?", "What evidence exists regarding 'joint 4'?", "What observation would rule out: Harmonic-drive backlash increased?", "What observation would rule out: Teach-pendant firmware corrupted joint 4 kinematics?"], "distractors": ["Harmonic-drive backlash increased", "Teach-pendant firmware corrupted joint 4 kinematics", "Laser tracker nest moved", "Bus-voltage sag caused missing encoder clocks", "encoder contamination"], "domain": "robotics", "evidence": "Messy dump distilled to:\nTITLE: Incomplete Robot Cell: Magnetic Encoder Drift After Nearby Grinding\nDOMAIN: robotics / joint_sensing\nOBJECTIVE: Keep a six-axis pick cell's tool-flange pose within 0.15 mm of the taught waypoint after a 4-hour production shift. Overnight, joint 4 began reporting a slow 0.8 deg drift and the cell started dropping pouches at the place nest.\nPROTOCOL:\n1. Home the robot and run the standard 12-point circularity check \u2014 Absolute magnetic encoder on joint 4, 19-bit, ring magnet bonded to the output shaft.\n2. Compare encoder counts to an independent laser tracker at the flange \u2014 Laser tracker is on a calibrated nest 1.8 m from the cell.\n3. Inspect joint 4 after the shift \u2014 No gearbox backlash change on the lock-to-lock test; grease looks normal.\nSENSOR_STATS:\n[pose_error] Tool-flange pose error versus the taught waypoint; obs: Error grows through the shift rather than appearing at power-up.; obs: The growth is almost entirely joint-4 yaw, not a base-frame shift.; n=6 mean=0.24 std=0.1241 min=0.08 max=0.41 mm RMS | [encoder_vs_tracker] Joint-4 angle: encoder minus laser-tracker reconstruction; obs: Disagreement peaks near the mid-travel zone, not at the hard stops.; obs: Repeating the same move without power cycling reproduces the hump.; n=10 mean=0.179 std=20 min=-40 max=40  | [gearbox_check] Mechanical health of joint 4; obs: Backlash, temperature, and bus voltage stay in-family.; n=4 mean=24.98 std=22.58 min=1.8 max=48.1 \nCONTEXT_HINTS:\n- A portable rail grinder was used on the neighboring cell for two hours that morning; sparks were visible over the shared aisle.\n- Joint 4's encoder is a magnetic ring with a vented cover, not an optical disk.\n- The teach pendant firmware was updated last month and has been stable on the other five joints.\n- Payload mass and TCP calibration were not changed.\n- The laser tracker battery was swapped at lunch; the afternoon survey still matches the morning tracker scale.", "failure_category": "reagent_material_flaw", "gold_cause": "Ferrous grinding swarf contaminated the magnetic encoder ring on joint 4. Local field distortion added a mid-travel count error that looked like kinematic drift. The mechanical gearbox is healthy; the withheld clue is the neighboring rail-grinder work over an unsealed magnetic encoder.\n", "gold_hypothesis_index": 0, "group_id": "yaml:RF-017", "id": "RF-017-AUG38", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Tool wear and dulling raise cutting force; undersize/smear often track tool condition and coolant."], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "joint_sensing", "tabular_stats": [0.24, 0.12409673645990855, 0.08, 0.41, 6.0, 0.179, 20.001611660063794, -40.0, 40.0, 10.0, 24.975, 22.582335463808874, 1.8, 48.1, 4.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Incomplete Robot Cell: Magnetic Encoder Drift After Nearby Grinding", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["storage fade", "The new electrolyte etches the NMC", "A wet glovebox poisoned only the new lot", "Formation was incomplete", "The scientific hypothesis \u2014 that post-gap fade is a cycle-life failure of the new electrolyte \u2014 is wrong. Both chemistries calendar-aged at 100% SOC and 40 C during the shutdown. The withheld oven record and the 30% SOC fridge control are the discriminating facts.\n", "Cycler current-sensor drift"], "difficulty": "hard", "discriminating_followups": ["What evidence exists regarding 'gap'?", "What evidence exists regarding 'both'?", "What evidence exists regarding 'LLI'?", "What observation would rule out: The new electrolyte etches the NMC?", "What observation would rule out: Cycler current-sensor drift?"], "distractors": ["The new electrolyte etches the NMC", "Cycler current-sensor drift", "A wet glovebox poisoned only the new lot", "Formation was incomplete", "storage fade"], "domain": "electrical", "evidence": "Post-cleanse extract for reasoning:\nTITLE: Incomplete Aging Study: Calendar Fade Misread as Cycle-Life Failure\nDOMAIN: energy / battery_aging\nGOAL: Show that a new electrolyte raises 25 C cycle life to 80% capacity at >600 cycles versus a 400-cycle baseline. After a summer storage gap, both chemistries sat at 42-45% capacity before the first post-gap cycle.\nSTEPS_FOLLOWED:\n1. Form cells and begin C/3 cycling at 25 C \u2014 Both new and baseline electrolytes reached cycle 80 near 96% capacity.\n2. Resume C/3 cycling after the shutdown \u2014 First discharge after storage is 43-46% for both electrolytes.\n3. Run dQ/dV and electrode harvest on one cell \u2014 Loss of lithium inventory dominates; transition-metal dissolution is minor.\nSENSOR_STATS:\n[capacity] Capacity through the shutdown; obs: Both chemistries lose ~50 points during the gap, then fade slowly again.; obs: The drop is not unique to the new electrolyte.; n=6 mean=61.95 std=24.15 min=44.1 max=96.2  | [dqdv] dQ/dV before and after the gap; obs: Graphite staging peaks lose area (LLI) with little peak shift that would indicate a bulk cathode crystal change.; obs: A harvested anode shows a thick SEI; cathode lattice parameters are in family. | [storage_record] Oven SOC and temperature during the shutdown (withheld); note=The oven SOC/temperature sheet was not attached to the aging report.; obs: 77 days at 40 C and 100% SOC is a known calendar-aging extreme.; n=4 mean=66.3 std=23.8 min=40.2 max=100 \nCLUES:\n- Cells sat 11 weeks at 100% SOC in a 40 C oven.\n- Calendar aging of NMC/graphite at high SOC is often larger than modest cycle counts.\n- The new electrolyte's selling point was cycle life, so the team framed every fade as a cycle-life failure.\n- A control that stayed at 30% SOC in a 10 C fridge over the same gap lost only 2%.\nDUMP_EXTRACT:\nEVIDENCE_PACKET epidebug.evidence_packet/v0.1\nSOURCE: /workspace/epidebug-space/repo/mock_data/dump_demo\nSUMMARIES:\n- NOTES.md: text lines=23 size=856\n- bore_finish.nc: text lines=5 size=54\n- charger.log: text lines=7 size=306\n- housing_revC.step: CAD step size=218\n- mill_cert_lot24081.txt: text lines=7 size=191\n- pack_temp.csv: CSV 5x3 schema={'time_s': 'numeric', 'temp_C': 'numeric', 'current_A': 'numeric'}\n- setup_vise.png: image [1, 1] mode=RGB\nANOMALY_CUES:\n- 2024-08-12 02:18:44 WARN cell_5 imbalance 42 mV\n- 2024-08-12 02:18:51 WARN cell_5 imbalance 48 mV\n- 2024-08-12 02:19:02 ERROR overheat pack_main\n- 2024-08-12 02:19:03 ALARM-401 thermal\n- 2024-08-12 02:19:04 ERROR overheat pack_main\nEXCERPTS:\n- [NOTES.md] Title: Housing bore drift / pack venting | Domain: machining | What went wrong: | The CNC bore is 0.18 mm undersize and the surface is smeared. After assembly, the last three 21700 cells in the 6S2P pack vented during a 2C charge. | Objective:\n- [bore_finish.nc] (REV C bore finish) | T4 M6 | S4200 M3 | G1 Z-12.0 F180 | M30\n- [charger.log] 2024-08-12 02:14:01 INFO charger ready | 2024-08-12 02:16:10 INFO CC 2.0C pack_main | 2024-08-12 02:18:44 WARN cell_5 imbalance 42 mV |", "failure_category": "flawed_hypothesis", "gold_cause": "The scientific hypothesis \u2014 that post-gap fade is a cycle-life failure of the new electrolyte \u2014 is wrong. Both chemistries calendar-aged at 100% SOC and 40 C during the shutdown. The withheld oven record and the 30% SOC fridge control are the discriminating facts.\n", "gold_hypothesis_index": 4, "group_id": "yaml:FH-016", "id": "FH-016-AUG692", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Check temperature sensor plausibility before accepting extreme readings as ground truth.", "Assay/identity of reagents and stoichiometry/mass balance before novel mechanism claims."], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "battery_aging", "tabular_stats": [61.949999999999996, 24.15220831863345, 44.1, 96.2, 6.0, 66.3, 23.802730935756088, 40.2, 100.0, 4.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.0, 1.4142135623730951, 0.0, 4.0, 5.0, 52.160000000000004, 31.96989834203418, 25.1, 94.0, 5.0, 2.02, 1.2140016474453401, 1.02, 3.6, 5.0, 0.0], "title": "Incomplete Aging Study: Calendar Fade Misread as Cycle-Life Failure", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Bearing defect progressing below vibration alarm threshold", "Machine failure label: L", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.0', 'Process temperature [K]': '310.5', 'Rotational speed [rpm]': '1263', 'Torque [Nm]': '60.6', 'Tool wear [min]': '90', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0027", "id": "HV-AI4I-0027", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4158.0, 302.0, 310.5, 1263.0, 60.6, 90.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 515.5083333333333, 0.0, 4158.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 27", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Power overload from high torque at high RPM rather than cooling deficit", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "Process physics consistent with a different AI4I failure mode than the label", "Tool wear failure: elevated torque from dull cutter misread as thermal trip"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Power overload from high torque at high RPM rather than cooling deficit", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Process physics consistent with a different AI4I failure mode than the label", "Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Process-temperature probe lag under-reporting while cooling is actually adequate"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '298.2', 'Process temperature [K]': '308.3', 'Rotational speed [rpm]': '1345', 'Torque [Nm]': '48.8', 'Tool wear [min]': '74', 'Failure Type': 'Heat Dissipation Failure', 'Machine failure': '1'}. Labeled fault class=Heat Dissipation Failure.", "failure_category": "reagent_material_flaw", "gold_cause": "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "gold_hypothesis_index": 3, "group_id": "harvest:HV-SYN-0009", "id": "HV-SYN-0009-X1-P2", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [298.2, 308.3, 1345.0, 48.8, 74.0, 1.0, 345.8833333333334, 1.0, 1345.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 9", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: M"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '300.6', 'Process temperature [K]': '309.4', 'Rotational speed [rpm]': '1367', 'Torque [Nm]': '63.2', 'Tool wear [min]': '45', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 5, "group_id": "harvest:HV-AI4I-0036", "id": "HV-AI4I-0036", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [2880.0, 300.6, 309.4, 1367.0, 63.2, 45.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 413.93333333333334, 0.0, 2880.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 36", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Machine failure label: L", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '303.6', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1222', 'Torque [Nm]': '66.8', 'Tool wear [min]': '148', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0007", "id": "HV-AI4I-0007", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4703.0, 303.6, 311.8, 1222.0, 66.8, 148.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 563.1, 0.0, 4703.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 7", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Machine failure label: L", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '298.0', 'Process temperature [K]': '308.1', 'Rotational speed [rpm]': '1454', 'Torque [Nm]': '37.1', 'Tool wear [min]': '35', 'Machine failure': '0'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 0, "group_id": "harvest:HV-AI4I-0098", "id": "HV-AI4I-0098", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [266.0, 298.0, 308.1, 1454.0, 37.1, 35.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 199.85, 0.0, 1454.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 98", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Machine failure label: L", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.4', 'Process temperature [K]': '311.9', 'Rotational speed [rpm]': '1298', 'Torque [Nm]': '63.9', 'Tool wear [min]': '59', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0060", "id": "HV-AI4I-0060-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4841.0, 303.4, 311.9, 1298.0, 63.9, 59.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 573.2666666666667, 0.0, 4841.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 60", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Column overloaded", "Blaming material when differential expansion or fixture compliance dominates", "Impurity truly present at 0.01%", "Noise fluctuation misread as peak", "Wrong integration baseline only"], "difficulty": "easy", "discriminating_followups": ["What is repeat-injection \u03c3?", "Is the peak above 3\u03c3 at the retention time?"], "distractors": ["Impurity truly present at 0.01%", "Column overloaded", "Wrong integration baseline only", "Blaming material when differential expansion or fixture compliance dominates"], "domain": "physics_priors", "evidence": "Claimed detection of a 0.01% impurity peak when chromatogram noise SD is 0.02% area and no repeat injections.", "failure_category": "flawed_hypothesis", "gold_cause": "Noise fluctuation misread as peak", "gold_hypothesis_index": 3, "group_id": "prior:PRIOR-011", "id": "PRIOR-011", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Reliable detection needs signal \u2273 3\u03c3 noise for a claim; 0.01 vs 0.02 fails.", "Single injection cannot separate noise fluke from analyte."], "source_file": null, "source_type": "synthetic_physics_prior", "split": "validation", "subdomain": "snr_detection", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Synthetic prior: snr_detection", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["Oven temperature was low", "Attributing missing mass/energy to new chemistry without closure checks", "Insufficient soak time vs diffusion timescale", "Substrate contamination only", "Resin lot is chemically off-spec"], "difficulty": "easy", "discriminating_followups": ["What is estimated D for the solvent/polymer pair?", "Was cross-section swelling measured vs time?"], "distractors": ["Resin lot is chemically off-spec", "Oven temperature was low", "Substrate contamination only", "Attributing missing mass/energy to new chemistry without closure checks"], "domain": "physics_priors", "evidence": "Team claims a 1 mm polymer coating fully solvent-equilibrated in 30 s at room temp after dip. Softening failure blamed on bad resin chemistry.", "failure_category": "flawed_hypothesis", "gold_cause": "Insufficient soak time vs diffusion timescale", "gold_hypothesis_index": 2, "group_id": "prior:PRIOR-008", "id": "PRIOR-008", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Diffusion time scale ~ L\u00b2/D; for L=1e-3 m and D~1e-12..1e-10 m\u00b2/s, t is minutes\u2013hours not seconds.", "30 s is implausibly short for full through-thickness equilibration."], "source_file": null, "source_type": "synthetic_physics_prior", "split": "validation", "subdomain": "diffusion_time", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Synthetic prior: diffusion_time", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Machine failure label: L", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '295.7', 'Process temperature [K]': '306.2', 'Rotational speed [rpm]': '2270', 'Torque [Nm]': '14.6', 'Tool wear [min]': '149', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 0, "group_id": "harvest:HV-AI4I-0031", "id": "HV-AI4I-0031", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [904.0, 295.7, 306.2, 2270.0, 14.6, 149.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 328.4583333333333, 0.0, 2270.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 31", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Cutting tool wear beyond process window as primary driver", "Process overstrain from recipe outside qualified envelope", "Power/torque path overload trip without recipe overstrain", "Feed-rate encoder fault reporting higher feed than actual", "Process physics consistent with a different AI4I failure mode than the label", "Heat dissipation failure from cooling shortfall at low RPM"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Cutting tool wear beyond process window as primary driver", "Heat dissipation failure from cooling shortfall at low RPM", "Power/torque path overload trip without recipe overstrain", "Feed-rate encoder fault reporting higher feed than actual", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '300.8', 'Process temperature [K]': '311.1', 'Rotational speed [rpm]': '1441', 'Torque [Nm]': '65.9', 'Tool wear [min]': '74', 'Failure Type': 'Overstrain Failure', 'Machine failure': '1'}. Failure type field=Overstrain Failure.", "failure_category": "protocol_human_loophole", "gold_cause": "Process overstrain from recipe outside qualified envelope", "gold_hypothesis_index": 1, "group_id": "harvest:HV-SYN-0012", "id": "HV-SYN-0012-X1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [300.8, 311.1, 1441.0, 65.9, 74.0, 1.0, 365.6333333333334, 1.0, 1441.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 12", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Current-transformer phase error inventing apparent power trip", "Power/torque path indicates overload trip condition", "Tool wear beyond process window elevating cutting torque", "Heat dissipation path obstructed at low rotational speed", "Process physics consistent with a different AI4I failure mode than the label", "Overstrain recipe outside qualified envelope without power-limit hit"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Current-transformer phase error inventing apparent power trip", "Process physics consistent with a different AI4I failure mode than the label", "Heat dissipation path obstructed at low rotational speed", "Overstrain recipe outside qualified envelope without power-limit hit", "Tool wear beyond process window elevating cutting torque"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '299.5', 'Process temperature [K]': '308.6', 'Rotational speed [rpm]': '1541', 'Torque [Nm]': '61.8', 'Tool wear [min]': '13', 'Failure Type': 'Power Failure', 'Machine failure': '1'}. Labeled fault class=Power Failure.", "failure_category": "instrumentation_artifact", "gold_cause": "Power/torque path indicates overload trip condition", "gold_hypothesis_index": 1, "group_id": "harvest:HV-SYN-0015", "id": "HV-SYN-0015-X1-P2", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [299.5, 308.6, 1541.0, 61.8, 13.0, 1.0, 370.81666666666666, 1.0, 1541.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 15", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Machine failure label: M", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial telemetry extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.1', 'Process temperature [K]': '310.7', 'Rotational speed [rpm]': '1376', 'Torque [Nm]': '46.3', 'Tool wear [min]': '210', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0057", "id": "HV-AI4I-0057-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4121.0, 302.1, 310.7, 1376.0, 46.3, 210.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 530.6750000000001, 0.0, 4121.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 57", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Machine failure label: M", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '300.6', 'Process temperature [K]': '309.4', 'Rotational speed [rpm]': '1367', 'Torque [Nm]': '63.2', 'Tool wear [min]': '45', 'Machine failure': '1'}. Labeled fault class=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0036", "id": "HV-AI4I-0036-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [2880.0, 300.6, 309.4, 1367.0, 63.2, 45.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 413.93333333333334, 0.0, 2880.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 36", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Machine failure label: L"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial sensor readings extract (AI4I-style). Key readings: {'Air temperature [K]': '302.3', 'Process temperature [K]': '311.6', 'Rotational speed [rpm]': '1488', 'Torque [Nm]': '41.8', 'Tool wear [min]': '75', 'Machine failure': '0'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 5, "group_id": "harvest:HV-AI4I-0080", "id": "HV-AI4I-0080-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3642.0, 302.3, 311.6, 1488.0, 41.8, 75.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 488.39166666666665, 0.0, 3642.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 80", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Machine failure label: L", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial telemetry extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.6', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1222', 'Torque [Nm]': '66.8', 'Tool wear [min]': '148', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0007", "id": "HV-AI4I-0007-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4703.0, 303.6, 311.8, 1222.0, 66.8, 148.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 563.1, 0.0, 4703.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 7", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["The LC-MS/MS method has autosampler/needle carryover after high-concentration standards. Residual analyte from the high standard appears in subsequent blanks and nearby samples, creating false low-level biomarker positives.\n", "Calibration curve is nonlinear", "Matrix effect suppresses internal standard", "sample carryover", "contaminated blanks", "True low-level biomarker is present in blanks"], "difficulty": "medium", "discriminating_followups": ["What observation would rule out: True low-level biomarker is present in blanks?", "What observation would rule out: Matrix effect suppresses internal standard?"], "distractors": ["True low-level biomarker is present in blanks", "Matrix effect suppresses internal standard", "Calibration curve is nonlinear", "sample carryover", "contaminated blanks"], "domain": "biology_lab", "evidence": "Messy dump distilled to:\nTITLE: LC-MS Carryover Creates False Low-Level Biomarker Positives\nDOMAIN: clinical / mass_spectrometry\nGOAL: Quantify a low-abundance drug metabolite in plasma samples by LC-MS/MS. Blank samples should be below the lower limit of quantification.\nPROTOCOL:\n1. Inject calibration standards from high to low concentration \u2014 Highest standard is 5000 ng/mL, followed by blanks and study samples.\n2. Run the study batch with the validated gradient \u2014 Autosampler needle wash uses 50% methanol.\n3. Review blanks and reinject after stronger needle wash \u2014 A 90% acetonitrile/isopropanol wash is tested after unexpected positives.\nSENSOR_STATS:\n[batch_sequence] Metabolite peak areas by injection position; obs: Blank peak area decays after the high standard.; obs: Low-positive samples are near the carryover tail.; n=6 mean=1.682e+05 std=3.631e+05 min=3900 max=9.8e+05  | [wash_test] Carryover after needle-wash change; obs: Strong organic wash pushes blank signal below LLOQ.; n=4 mean=3855 std=3288 min=220 max=9200 \nCLUES:\n- The suspected biomarker is sticky and hydrophobic.\n- The high standard was injected before blanks rather than after blanks.\n- Retention time and product-ion ratio in blanks match the high standard.\n- Patient samples with positives cluster after high-concentration injections.", "failure_category": "instrumentation_artifact", "gold_cause": "The LC-MS/MS method has autosampler/needle carryover after high-concentration standards. Residual analyte from the high standard appears in subsequent blanks and nearby samples, creating false low-level biomarker positives.\n", "gold_hypothesis_index": 0, "group_id": "yaml:IN-013", "id": "IN-013-AUG936", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "mass_spectrometry", "tabular_stats": [168183.33333333334, 363059.31295343774, 3900.0, 980000.0, 6.0, 3855.0, 3288.0199208642275, 220.0, 9200.0, 4.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "LC-MS Carryover Creates False Low-Level Biomarker Positives", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Machine failure label: M", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.2', 'Process temperature [K]': '311.5', 'Rotational speed [rpm]': '1734', 'Torque [Nm]': '25.0', 'Tool wear [min]': '116', 'Machine failure': '0'}. Labeled fault class=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0093", "id": "HV-AI4I-0093-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3660.0, 302.2, 311.5, 1734.0, 25.0, 116.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 512.3916666666667, 0.0, 3660.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 93", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Machine failure label: L", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Shop-floor measurement digest (AI4I-style). Key readings: {'Air temperature [K]': '298.0', 'Process temperature [K]': '308.1', 'Rotational speed [rpm]': '1454', 'Torque [Nm]': '37.1', 'Tool wear [min]': '35', 'Machine failure': '0'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 0, "group_id": "harvest:HV-AI4I-0098", "id": "HV-AI4I-0098-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [266.0, 298.0, 308.1, 1454.0, 37.1, 35.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 199.85, 0.0, 1454.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 98", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Tool wear beyond process window elevating cutting torque", "Heat dissipation path obstructed at low rotational speed", "Power/torque path indicates overload trip condition", "Overstrain recipe outside qualified envelope without power-limit hit", "Process physics consistent with a different AI4I failure mode than the label", "Current-transformer phase error inventing apparent power trip"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Heat dissipation path obstructed at low rotational speed", "Tool wear beyond process window elevating cutting torque", "Overstrain recipe outside qualified envelope without power-limit hit", "Current-transformer phase error inventing apparent power trip", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '298.1', 'Process temperature [K]': '308.7', 'Rotational speed [rpm]': '1629', 'Torque [Nm]': '62.6', 'Tool wear [min]': '47', 'Failure Type': 'Power Failure', 'Machine failure': '1'}. Failure type field=Power Failure.", "failure_category": "instrumentation_artifact", "gold_cause": "Power/torque path indicates overload trip condition", "gold_hypothesis_index": 2, "group_id": "harvest:HV-SYN-0003", "id": "HV-SYN-0003-X1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [298.1, 308.7, 1629.0, 62.6, 47.0, 1.0, 391.06666666666666, 1.0, 1629.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 3", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: M", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '303.4', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1316', 'Torque [Nm]': '50.9', 'Tool wear [min]': '69', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0041", "id": "HV-AI4I-0041", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4845.0, 303.4, 311.8, 1316.0, 50.9, 69.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 574.8416666666667, 0.0, 4845.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 41", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Feed-rate encoder fault reporting higher feed than actual", "Heat dissipation failure from cooling shortfall at low RPM", "Power/torque path overload trip without recipe overstrain", "Process overstrain from recipe outside qualified envelope", "Process physics consistent with a different AI4I failure mode than the label", "Cutting tool wear beyond process window as primary driver"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Heat dissipation failure from cooling shortfall at low RPM", "Feed-rate encoder fault reporting higher feed than actual", "Power/torque path overload trip without recipe overstrain", "Cutting tool wear beyond process window as primary driver", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '300.8', 'Process temperature [K]': '311.1', 'Rotational speed [rpm]': '1441', 'Torque [Nm]': '65.9', 'Tool wear [min]': '74', 'Failure Type': 'Overstrain Failure', 'Machine failure': '1'}. Labeled fault class=Overstrain Failure.", "failure_category": "protocol_human_loophole", "gold_cause": "Process overstrain from recipe outside qualified envelope", "gold_hypothesis_index": 3, "group_id": "harvest:HV-SYN-0012", "id": "HV-SYN-0012-X1-P2", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [300.8, 311.1, 1441.0, 65.9, 74.0, 1.0, 365.6333333333334, 1.0, 1441.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 12", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Machine failure label: L", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Shop-floor measurement digest (AI4I-style). Salient channels: {'Air temperature [K]': '301.8', 'Process temperature [K]': '310.2', 'Rotational speed [rpm]': '2825', 'Torque [Nm]': '5.8', 'Tool wear [min]': '215', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 0, "group_id": "harvest:HV-AI4I-0002", "id": "HV-AI4I-0002-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4297.0, 301.8, 310.2, 2825.0, 5.8, 215.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 663.0666666666667, 0.0, 4297.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 2", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Machine failure label: L"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.0', 'Process temperature [K]': '310.5', 'Rotational speed [rpm]': '1263', 'Torque [Nm]': '60.6', 'Tool wear [min]': '90', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 5, "group_id": "harvest:HV-AI4I-0027", "id": "HV-AI4I-0027-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4158.0, 302.0, 310.5, 1263.0, 60.6, 90.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 515.5083333333333, 0.0, 4158.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 27", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Machine failure label: L", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.6', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1222', 'Torque [Nm]': '66.8', 'Tool wear [min]': '148', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0007", "id": "HV-AI4I-0007-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4703.0, 303.6, 311.8, 1222.0, 66.8, 148.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 563.1, 0.0, 4703.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 7", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: L", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.3', 'Process temperature [K]': '311.4', 'Rotational speed [rpm]': '1333', 'Torque [Nm]': '66.7', 'Tool wear [min]': '205', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0053", "id": "HV-AI4I-0053", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3944.0, 302.3, 311.4, 1333.0, 66.7, 205.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0, 513.7833333333333, 0.0, 3944.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 53", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Machine failure label: L", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.0', 'Process temperature [K]': '310.5', 'Rotational speed [rpm]': '1263', 'Torque [Nm]': '60.6', 'Tool wear [min]': '90', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0027", "id": "HV-AI4I-0027-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4158.0, 302.0, 310.5, 1263.0, 60.6, 90.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 515.5083333333333, 0.0, 4158.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 27", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Machine failure label: L", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Industrial telemetry extract (AI4I-style). Salient channels: {'Air temperature [K]': '298.0', 'Process temperature [K]': '308.1', 'Rotational speed [rpm]': '1454', 'Torque [Nm]': '37.1', 'Tool wear [min]': '35', 'Machine failure': '0'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 0, "group_id": "harvest:HV-AI4I-0098", "id": "HV-AI4I-0098-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [266.0, 298.0, 308.1, 1454.0, 37.1, 35.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 199.85, 0.0, 1454.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 98", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Machine failure label: M", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Shop-floor measurement digest (AI4I-style). Salient channels: {'Air temperature [K]': '302.4', 'Process temperature [K]': '310.3', 'Rotational speed [rpm]': '1378', 'Torque [Nm]': '54.3', 'Tool wear [min]': '165', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 0, "group_id": "harvest:HV-AI4I-0015", "id": "HV-AI4I-0015-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4450.0, 302.4, 310.3, 1378.0, 54.3, 165.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 555.1666666666666, 0.0, 4450.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 15", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Machine failure label: M", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.2', 'Process temperature [K]': '311.5', 'Rotational speed [rpm]': '1734', 'Torque [Nm]': '25.0', 'Tool wear [min]': '116', 'Machine failure': '0'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0093", "id": "HV-AI4I-0093-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3660.0, 302.2, 311.5, 1734.0, 25.0, 116.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 512.3916666666667, 0.0, 3660.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 93", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Machine failure label: L", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.3', 'Process temperature [K]': '311.6', 'Rotational speed [rpm]': '1488', 'Torque [Nm]': '41.8', 'Tool wear [min]': '75', 'Machine failure': '0'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0080", "id": "HV-AI4I-0080", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3642.0, 302.3, 311.6, 1488.0, 41.8, 75.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 488.39166666666665, 0.0, 3642.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 80", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Machine failure label: M", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '300.6', 'Process temperature [K]': '309.4', 'Rotational speed [rpm]': '1367', 'Torque [Nm]': '63.2', 'Tool wear [min]': '45', 'Machine failure': '1'}. Labeled fault class=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0036", "id": "HV-AI4I-0036-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [2880.0, 300.6, 309.4, 1367.0, 63.2, 45.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 413.93333333333334, 0.0, 2880.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 36", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["True mechanical resonance at 500 Hz", "Software windowing artifact only", "Aliasing from higher-frequency source / missing anti-alias filter", "Loose accelerometer mount only", "Unit scaling mismatch between transducer and PLC/engineering units"], "difficulty": "easy", "discriminating_followups": ["What anti-alias filter is in the DAQ chain?", "Does a higher sample rate move the peak?"], "distractors": ["True mechanical resonance at 500 Hz", "Loose accelerometer mount only", "Software windowing artifact only", "Unit scaling mismatch between transducer and PLC/engineering units"], "domain": "physics_priors", "evidence": "Vibration FFT shows a strong peak at exactly 50% of the sampling rate. Mechanical inspection finds no corresponding resonance. Sample rate 1 kHz.", "failure_category": "instrumentation_artifact", "gold_cause": "Aliasing from higher-frequency source / missing anti-alias filter", "gold_hypothesis_index": 2, "group_id": "prior:PRIOR-006", "id": "PRIOR-006", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Nyquist: content above fs/2 aliases.", "Peak glued to Nyquist often indicates aliasing or anti-alias filter failure."], "source_file": null, "source_type": "synthetic_physics_prior", "split": "validation", "subdomain": "sampling_aliasing", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Synthetic prior: sampling_aliasing", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Charger current sensor failed instead", "Ambient lab fire", "Thermistor wiring short / ADC saturation artifact", "Aliasing / missing anti-alias filter inventing a false spectral peak", "True thermal runaway already at 950 C"], "difficulty": "easy", "discriminating_followups": ["Does an IR gun on the can agree?", "Is the thermistor resistance in-range at room temp?"], "distractors": ["True thermal runaway already at 950 C", "Charger current sensor failed instead", "Ambient lab fire", "Aliasing / missing anti-alias filter inventing a false spectral peak"], "domain": "physics_priors", "evidence": "Thermistor on pack can reports 950 C during a 2C charge of a Li-ion module while current log looks normal and no venting observed yet in that second.", "failure_category": "instrumentation_artifact", "gold_cause": "Thermistor wiring short / ADC saturation artifact", "gold_hypothesis_index": 2, "group_id": "prior:PRIOR-003", "id": "PRIOR-003", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Li-ion pack can temperatures >200 C imply thermal runaway; 950 C is near steel glow \u2014 check sensor first.", "Single-point thermistor taped to can can short or float."], "source_file": null, "source_type": "synthetic_physics_prior", "split": "validation", "subdomain": "sensor_plausibility", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Synthetic prior: sensor_plausibility", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: M", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.1', 'Process temperature [K]': '310.7', 'Rotational speed [rpm]': '1376', 'Torque [Nm]': '46.3', 'Tool wear [min]': '210', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0057", "id": "HV-AI4I-0057", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4121.0, 302.1, 310.7, 1376.0, 46.3, 210.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 530.6750000000001, 0.0, 4121.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 57", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Oven thermocouple failed open", "Anti-windup disabled after actuator saturation", "Atmosphere was oxidizing due to door seal leak only", "Material lot was contaminated", "Fahrenheit/Celsius unit confusion on setpoint"], "difficulty": "easy", "discriminating_followups": ["Which unit does the controller UI display?", "What was the independent pyrometer reading?"], "distractors": ["Material lot was contaminated", "Oven thermocouple failed open", "Atmosphere was oxidizing due to door seal leak only", "Anti-windup disabled after actuator saturation"], "domain": "physics_priors", "evidence": "Protocol says bake at 350 F. Logged controller setpoint is 350 C. Coupon oxidized heavily.", "failure_category": "protocol_human_loophole", "gold_cause": "Fahrenheit/Celsius unit confusion on setpoint", "gold_hypothesis_index": 4, "group_id": "prior:PRIOR-002", "id": "PRIOR-002", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["350 F \u2248 176.7 C; 350 C \u2248 662 F.", "Unit mismatch can shift thermal exposure by hundreds of degrees."], "source_file": null, "source_type": "synthetic_physics_prior", "split": "validation", "subdomain": "units_temperature", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Synthetic prior: units_temperature", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Machine failure label: M", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.4', 'Process temperature [K]': '310.3', 'Rotational speed [rpm]': '1378', 'Torque [Nm]': '54.3', 'Tool wear [min]': '165', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0015", "id": "HV-AI4I-0015", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4450.0, 302.4, 310.3, 1378.0, 54.3, 165.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 555.1666666666666, 0.0, 4450.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 15", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Machine failure label: M", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.0', 'Process temperature [K]': '309.7', 'Rotational speed [rpm]': '1386', 'Torque [Nm]': '62.7', 'Tool wear [min]': '142', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0019", "id": "HV-AI4I-0019", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4354.0, 302.0, 309.7, 1386.0, 62.7, 142.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 546.5333333333333, 0.0, 4354.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 19", "why_alternatives_fail": null}
{"category_label": "flawed_hypothesis", "causal_chain": [], "competing_hypotheses": ["The new substrate lot lowered gain", "Horn standard is 4 dB off", "range too short", "Chamber multipath", "The measurement hypothesis \u2014 that received power follows inverse-square Friis at 0.42 m \u2014 is invalid in the radiative near field. Aperture phase curvature makes apparent gain range-dependent. The antenna is ~6.5 dBi once a far-field distance or a near-field transform is used.\n", "Uncorrected cable loss"], "difficulty": "hard", "discriminating_followups": ["What evidence exists regarding '0.42'?", "What evidence exists regarding '0.53'?", "What evidence exists regarding 'phase'?", "What observation would rule out: The new substrate lot lowered gain?", "What observation would rule out: Uncorrected cable loss?"], "distractors": ["The new substrate lot lowered gain", "Uncorrected cable loss", "Chamber multipath", "Horn standard is 4 dB off", "range too short"], "domain": "sensors", "evidence": "Post-cleanse extract for reasoning:\nTITLE: Antenna Gain Error from Applying Inverse-Square in the Near Field\nDOMAIN: physics / rf_metrology\nGOAL: Measure 2.45 GHz patch-antenna gain in an anechoic corner to +/-0.6 dB using a 10 dBi standard horn. The team used Friis / inverse-square at 0.42 m separation and reported 2.1 dBi instead of the expected 6.5 dBi.\nPROTOCOL:\n1. Place the AUT and horn 0.42 m apart, boresight aligned \u2014 AUT aperture D = 0.18 m; far-field criterion 2D^2/lambda is 0.53 m. They used 0.42 m 'because the chamber is small'.\n2. Record S21 and convert to gain with Friis \u2014 Cable and mismatch corrections are applied; they are not the 4 dB miss.\n3. Compare to a 1.10 m far-field point and a compact-range borrow \u2014 At 1.10 m the same Friis reduction gives 6.4 dBi; a compact range gives 6.6 dBi.\nSENSOR_STATS:\n[gain_vs_range] Friis-derived apparent gain versus separation; obs: Apparent gain is a strong function of r below 0.8 m.; obs: It asymptotes near 6.4 dBi once r exceeds ~1.5 x 2D^2/lambda.; n=11 mean=1.985 std=2.372 min=-0.8 max=6.4  | [phase] Aperture phase flatness from a near-field probe; obs: At 0.42 m the AUT aperture phase varies 38 deg peak-to-peak.; obs: At 1.10 m the variation is 9 deg.; obs: A spherical-wave / 1/r^2 assumption is invalid with that curvature. | [rf_integrity] Cable, mismatch, and chamber checks; obs: Mismatch and cable terms are already in the reduction.; obs: Chamber reflectivity cannot explain a smooth 4 dB range dependence.; n=4 mean=-18.45 std=14.07 min=-38 max=1.12 \nCONTEXT_HINTS:\n- 2D^2/lambda at 2.45 GHz for D = 0.18 m is 0.53 m; they measured at 0.42 m.\n- Friis assumes plane-wave / far-field incidence.\n- The chamber length limited the first setup.\n- A student argued the patch 'just has low gain' because of a new substrate lot.", "failure_category": "flawed_hypothesis", "gold_cause": "The measurement hypothesis \u2014 that received power follows inverse-square Friis at 0.42 m \u2014 is invalid in the radiative near field. Aperture phase curvature makes apparent gain range-dependent. The antenna is ~6.5 dBi once a far-field distance or a near-field transform is used.\n", "gold_hypothesis_index": 4, "group_id": "yaml:FH-021", "id": "FH-021-AUG354", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": [], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "rf_metrology", "tabular_stats": [1.9845454545454546, 2.372342009822835, -0.8, 6.4, 11.0, -18.445, 14.069757460596113, -38.0, 1.12, 4.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Antenna Gain Error from Applying Inverse-Square in the Near Field", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Machine failure label: M", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.4', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1316', 'Torque [Nm]': '50.9', 'Tool wear [min]': '69', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0041", "id": "HV-AI4I-0041-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4845.0, 303.4, 311.8, 1316.0, 50.9, 69.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 574.8416666666667, 0.0, 4845.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 41", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: M", "Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.7', 'Process temperature [K]': '312', 'Rotational speed [rpm]': '2663', 'Torque [Nm]': '11.4', 'Tool wear [min]': '71', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0062", "id": "HV-AI4I-0062-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4763.0, 303.7, 312.0, 2663.0, 11.4, 71.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 677.1750000000001, 0.0, 4763.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 62", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["The 4340 heat is off-chemistry", "Quench oil was too hot or watery", "Severe decarburization", "Temper overshot and softened everything", "The operator double-stacked 48 blocks and started the 40-minute soak when the air thermocouple hit set-point, not when the heaviest part did. Center cores never reached austenitize temperature, so they formed mixed bainite/pearlite and finished soft. Oil, alloy, and temper are secondary.\n", "under-soaked load"], "difficulty": "medium", "discriminating_followups": ["What evidence exists regarding 'center'?", "What evidence exists regarding '801'?", "What evidence exists regarding 'basket'?", "What observation would rule out: Quench oil was too hot or watery?", "What observation would rule out: The 4340 heat is off-chemistry?"], "distractors": ["Quench oil was too hot or watery", "The 4340 heat is off-chemistry", "Temper overshot and softened everything", "Severe decarburization", "under-soaked load"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse extract for reasoning:\nTITLE: Soft Cores from a Heat-Treat Load Packed Too Densely\nDOMAIN: manufacturing / heat_treatment\nOBJECTIVE: Through-harden 4340 coupon blocks to 50-53 HRC with a 845 C soak and oil quench. Surface and mid-thickness hardness should agree within 2 HRC.\nPROTOCOL:\n1. Load 48 blocks into a batch furnace basket \u2014 SOP allows 24 blocks with 15 mm spacing. The night shift double-stacked to finish the order.\n2. Soak 40 minutes from the air-TC set-point, then oil quench \u2014 SOP soak is 40 minutes after the heaviest part is within 10 C of set-point.\n3. Temper 200 C / 2 h and hardness-map six blocks from edge, middle, and center of the basket \u2014 Edge blocks pass; center blocks are 42-46 HRC in the core.\nTELEMETRY_EXTRACT:\n[hardness] HRC after temper, surface versus mid-plane; obs: Basket-edge parts are through-hard.; obs: Basket-center cores are 6-8 HRC soft.; n=6 mean=48.73 std=3.041 min=43.3 max=51.8  | [furnace] Air thermocouple and a reconstructed load estimate; obs: Oil temperature is in family.; obs: A later instrumented dummy in the same packing peaks at 801 C when the air TC is already at 845 C.; n=4 mean=437 std=386.4 min=40 max=845  | [microstructure] Nital etch of a center-core coupon; obs: Mixed martensite plus fine pearlite/bainite in the core.; obs: Edge coupons are nearly full martensite.; obs: No decarburization deeper than 0.04 mm.\nCONTEXT_HINTS:\n- The SOP sketch shows a single layer with 15 mm gaps.\n- The same 4340 heat through-hardened last week in a 24-piece load.\n- Quench oil was analyzed last month and is within water and soot limits.\n- A new temper furnace was commissioned, so some blamed temper color.\nDUMP_EXTRACT:\nEVIDENCE_PACKET epidebug.evidence_packet/v0.1\nSOURCE: /workspace/epidebug-space/repo/mock_data/dump_demo\nSUMMARIES:\n- NOTES.md: text lines=23 size=856\n- bore_finish.nc: text lines=5 size=54\n- charger.log: text lines=7 size=306\n- housing_revC.step: CAD step size=218\n- mill_cert_lot24081.txt: text lines=7 size=191\n- pack_temp.csv: CSV 5x3 schema={'time_s': 'numeric', 'temp_C': 'numeric', 'current_A': 'numeric'}\n- setup_vise.png: image [1, 1] mode=RGB\nANOMALY_CUES:\n- 2024-08-12 02:18:44 WARN cell_5 imbalance 42 mV\n- 2024-08-12 02:18:51 WARN cell_5 imbalance 48 mV\n- 2024-08-12 02:19:02 ERROR overheat pack_main\n- 2024-08-12 02:19:03 ALARM-401 thermal\n- 2024-08-12 02:19:04 ERROR overheat pack_main\nEXCERPTS:\n- [NOTES.md] Title: Housing bore drift / pack venting | Domain: machining | What went wrong: | The CNC bore is 0.18 mm undersize and the surface is smeared. After assembly, the last three 21700 cells in the 6S2P pack vented during a 2C charge. | Objective:\n- [bore_finish.nc] (REV C bore finish) | T4 M6 | S4200 M3 | G1 Z-12.0 F180 | M30\n- [charger.log] 2024-08-12 02:14:01 INFO charger ready | 2024-08-12 02:16:10 INFO CC 2.0C pack_main | 2024-08-12 02:18:44 WARN cell_5 imbalance 42 mV |", "failure_category": "protocol_human_loophole", "gold_cause": "The operator double-stacked 48 blocks and started the 40-minute soak when the air thermocouple hit set-point, not when the heaviest part did. Center cores never reached austenitize temperature, so they formed mixed bainite/pearlite and finished soft. Oil, alloy, and temper are secondary.\n", "gold_hypothesis_index": 4, "group_id": "yaml:PL-011", "id": "PL-011-AUG68", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Check temperature sensor plausibility before accepting extreme readings as ground truth."], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "heat_treatment", "tabular_stats": [48.73333333333333, 3.04065051519499, 43.3, 51.8, 6.0, 437.0, 386.39164069632767, 40.0, 845.0, 4.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.0, 1.4142135623730951, 0.0, 4.0, 5.0, 52.160000000000004, 31.96989834203418, 25.1, 94.0, 5.0, 2.02, 1.2140016474453401, 1.02, 3.6, 5.0, 0.0], "title": "Soft Cores from a Heat-Treat Load Packed Too Densely", "why_alternatives_fail": null}
{"category_label": "protocol_human_loophole", "causal_chain": [], "competing_hypotheses": ["Heat dissipation failure from cooling shortfall at low RPM", "Power/torque path overload trip without recipe overstrain", "Process overstrain from recipe outside qualified envelope", "Process physics consistent with a different AI4I failure mode than the label", "Feed-rate encoder fault reporting higher feed than actual", "Cutting tool wear beyond process window as primary driver"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Heat dissipation failure from cooling shortfall at low RPM", "Process physics consistent with a different AI4I failure mode than the label", "Power/torque path overload trip without recipe overstrain", "Cutting tool wear beyond process window as primary driver", "Feed-rate encoder fault reporting higher feed than actual"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '300.8', 'Process temperature [K]': '311.1', 'Rotational speed [rpm]': '1441', 'Torque [Nm]': '65.9', 'Tool wear [min]': '74', 'Failure Type': 'Overstrain Failure', 'Machine failure': '1'}. Failure type field=Overstrain Failure.", "failure_category": "protocol_human_loophole", "gold_cause": "Process overstrain from recipe outside qualified envelope", "gold_hypothesis_index": 2, "group_id": "harvest:HV-SYN-0012", "id": "HV-SYN-0012-X1-P1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [300.8, 311.1, 1441.0, 65.9, 74.0, 1.0, 365.6333333333334, 1.0, 1441.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 12", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Machine failure label: M", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.7', 'Process temperature [K]': '312', 'Rotational speed [rpm]': '2663', 'Torque [Nm]': '11.4', 'Tool wear [min]': '71', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0062", "id": "HV-AI4I-0062-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4763.0, 303.7, 312.0, 2663.0, 11.4, 71.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 677.1750000000001, 0.0, 4763.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 62", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["probe fouling", "dirty sensor", "Construction runoff event", "Benchtop turbidimeter failed", "The turbidity spike is an optical measurement artifact caused by biofouling and silt on the probe window. The river did not become turbid; the dirty window scattered light and inflated the NTU reading.\n", "Natural storm-driven sediment pulse"], "difficulty": "easy", "discriminating_followups": ["What observation would rule out: Construction runoff event?", "What observation would rule out: Benchtop turbidimeter failed?"], "distractors": ["Construction runoff event", "Benchtop turbidimeter failed", "Natural storm-driven sediment pulse", "dirty sensor", "probe fouling"], "domain": "sensors", "evidence": "Post-cleanse extract for reasoning:\nTITLE: River Turbidity Spike from Biofouled Optical Sensor Window\nDOMAIN: environmental_science / water_quality_monitoring\nOBJECTIVE: Monitor river turbidity downstream of a construction site. Regulatory alerts trigger if turbidity exceeds baseline by more than 50 NTU for over 30 minutes.\nPROTOCOL:\n1. Deploy optical turbidity probe on a fixed river platform \u2014 Probe reports 5-minute NTU averages and is cleaned every two weeks.\n2. Investigate a sudden turbidity alert \u2014 The alert occurs after 12 days without rain.\n3. Collect grab samples and inspect the probe window \u2014 Manual samples are measured on a benchtop turbidimeter.\nSENSOR_STATS:\n[field_probe] Field turbidity sensor readings; obs: Probe reports a sharp turbidity increase without rainfall or flow increase.; n=8 mean=64.75 std=35.94 min=18 max=102 NTU | [grab_samples] Manual grab-sample turbidity; obs: Manual samples remain near baseline.; n=3 mean=20.33 std=1.247 min=19 max=22 NTU | [inspection] Probe inspection; obs: Optical window is covered by algae film and trapped silt.; obs: After cleaning, field probe immediately reads 21 NTU.\nCONTEXT_HINTS:\n- Water temperature increased during the week before the alert.\n- No rainfall, excavation, or upstream flow change is recorded.\n- The cleaning interval was extended because staff were unavailable.\n- A second upstream station did not show a turbidity spike.", "failure_category": "instrumentation_artifact", "gold_cause": "The turbidity spike is an optical measurement artifact caused by biofouling and silt on the probe window. The river did not become turbid; the dirty window scattered light and inflated the NTU reading.\n", "gold_hypothesis_index": 4, "group_id": "yaml:IN-014", "id": "IN-014-AUG26", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Instrumentation artifacts (calibration, aliasing, grounding) compete with process faults."], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "water_quality_monitoring", "tabular_stats": [64.75, 35.94353210245203, 18.0, 102.0, 8.0, 20.333333333333332, 1.247219128924647, 19.0, 22.0, 3.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "River Turbidity Spike from Biofouled Optical Sensor Window", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Machine failure label: M", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Shop-floor measurement digest (AI4I-style). Salient channels: {'Air temperature [K]': '302.1', 'Process temperature [K]': '310.7', 'Rotational speed [rpm]': '1376', 'Torque [Nm]': '46.3', 'Tool wear [min]': '210', 'Machine failure': '1'}. Labeled fault class=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0057", "id": "HV-AI4I-0057-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4121.0, 302.1, 310.7, 1376.0, 46.3, 210.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 530.6750000000001, 0.0, 4121.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 57", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Machine failure label: L", "Bearing defect progressing below vibration alarm threshold"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '295.7', 'Process temperature [K]': '306.2', 'Rotational speed [rpm]': '2270', 'Torque [Nm]': '14.6', 'Tool wear [min]': '149', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0031", "id": "HV-AI4I-0031-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [904.0, 295.7, 306.2, 2270.0, 14.6, 149.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 328.4583333333333, 0.0, 2270.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 31", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Current-transformer phase error inventing apparent power trip", "Overstrain recipe outside qualified envelope without power-limit hit", "Tool wear beyond process window elevating cutting torque", "Power/torque path indicates overload trip condition", "Heat dissipation path obstructed at low rotational speed", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Heat dissipation path obstructed at low rotational speed", "Process physics consistent with a different AI4I failure mode than the label", "Overstrain recipe outside qualified envelope without power-limit hit", "Current-transformer phase error inventing apparent power trip", "Tool wear beyond process window elevating cutting torque"], "domain": "manufacturing_industrial", "evidence": "Operator dump summary after cleanse: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '299.5', 'Process temperature [K]': '308.6', 'Rotational speed [rpm]': '1541', 'Torque [Nm]': '61.8', 'Tool wear [min]': '13', 'Failure Type': 'Power Failure', 'Machine failure': '1'}. Labeled fault class=Power Failure.", "failure_category": "instrumentation_artifact", "gold_cause": "Power/torque path indicates overload trip condition", "gold_hypothesis_index": 3, "group_id": "harvest:HV-SYN-0015", "id": "HV-SYN-0015-X1-P1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [299.5, 308.6, 1541.0, 61.8, 13.0, 1.0, 370.81666666666666, 1.0, 1541.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 15", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Machine failure label: L", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '301.8', 'Process temperature [K]': '310.2', 'Rotational speed [rpm]': '2825', 'Torque [Nm]': '5.8', 'Tool wear [min]': '215', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0002", "id": "HV-AI4I-0002", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4297.0, 301.8, 310.2, 2825.0, 5.8, 215.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 663.0666666666667, 0.0, 4297.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 2", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Material stock hardness was out of specification", "The CNC program used the wrong cutter compensation", "Fixture shifted during the run", "Coolant concentration was too low", "The reused 6 mm carbide end mill was worn beyond its tool-life limit. Flank wear and minor flute chipping increased cutting forces, vibration, and tool deflection during slot finishing. The worn cutter rubbed instead of shearing cleanly, creating burrs, poor surface finish, and undersized slots in later parts.\n", "worn cutter"], "difficulty": "medium", "discriminating_followups": ["What observation would rule out: Material stock hardness was out of specification?", "What observation would rule out: Fixture shifted during the run?"], "distractors": ["Material stock hardness was out of specification", "Fixture shifted during the run", "Coolant concentration was too low", "The CNC program used the wrong cutter compensation", "worn cutter"], "domain": "manufacturing_industrial", "evidence": "Cleaned evidence pack:\nTITLE: CNC End Mill Wear Drives Dimensional Drift and Burr Formation\nDOMAIN: manufacturing / cnc_milling\nGOAL: Machine a batch of aluminum 6061-T6 sensor brackets with a 6.00 +/- 0.03 mm mounting slot, surface roughness Ra below 1.6 um, and no visible burrs. The first 80 parts passed inspection, but parts from the second half of the batch failed dimensional and surface-finish acceptance.\nSTEPS_FOLLOWED:\n1. Load certified 6061-T6 plate stock into the Haas VF-2 CNC mill \u2014 Material certificate lists yield strength 276 MPa and hardness inside expected range.\n2. Install 6 mm three-flute carbide end mill for slot finishing \u2014 Tool ID EM-6-F3 was reused from the previous production run.\n3. Run the validated slot-finishing toolpath \u2014 Spindle 12000 rpm, feed 950 mm/min, 0.4 mm radial engagement, flood coolant.\n4. Inspect every twentieth part on the CMM and surface profilometer \u2014 Slot width, burr height, and Ra are measured after deburring is intentionally skipped for process validation.\n5. Inspect the end mill under a toolmaker microscope after the failed batch \u2014 Measure flank wear land and check cutting-edge chipping.\nSENSOR_STATS:\n[dimensional_trend] CMM slot-width measurements across the batch; obs: Slot width drifts below the lower tolerance after approximately part 100.; obs: The drift is gradual and monotonic rather than random.; n=18 mean=43.31 std=51.74 min=5.88 max=160  | [surface_finish] Surface roughness and burr-height measurements; obs: Ra exceeds the 1.6 um limit beginning near part 100.; obs: Burr height increases with surface roughness.; n=15 mean=32.91 std=53.54 min=0.01 max=160  | [machine_load] Spindle load and vibration trend from machine telemetry; obs: Spindle load and vibration rise as the batch progresses.; obs: Coolant flow and spindle speed remain stable.; n=24 mean=47.09 std=45.93 min=0.18 max=160  | [tool_inspection] Microscope inspection of tool EM-6-F3 after the run; obs: The end mill is more than twice its recommended tool-life.; obs: Measured flank wear exceeds the replacement threshold.; n=5 mean=5.588 std=7.037 min=0.2 max=18.4 \nCLUES:\n- The same plate stock lot produced both passing early parts and failing late parts.\n- Coolant concentration was 7.8%, inside the 7-9% process window.\n- The fixture was probed before and after the run; work offset shifted by less than 0.005 mm.\n- The tool had been used for an earlier abrasive prototype run but the operator's note was not copied into the production traveler.\n- The CNC program hash matches the previously validated process plan.\nANOMALY_CUES:\n- Slot width drifts below the lower tolerance after approximately part 100.\n- The drift is gradual and monotonic rather than random.\nDUMP_EXTRACT:\nEVIDENCE_PACKET epidebug.evidence_packet/v0.1\nSOURCE: /workspace/epidebug-space/repo/mock_data/dump_demo\nSUMMARIES:\n- NOTES.md: text lines=23 size=856\n- bore_finish.nc: text lines=5 size=54\n- charger.log: text lines=7 size=306\n- housing_revC.step: CAD step size=218\n- mill_cert_lot24081.txt: text lines=7 size=191\n- pack_temp.csv: CSV 5x3 schema={'time_s': 'numeric', 'temp_C': 'numeric', 'current_A': 'numeric'}\n- setup_vise.png: image [1, 1] mode=RGB\nANOMALY_CUES:\n- 2024-08-12 02:18:44 WARN cell_5 imbalance 42 mV\n- 2024-08-12 02:18:51 WARN cell_5 imbalance 48 mV\n- 2024-08-12 02:19:02 ERROR overheat pack_main\n- 2024-08-12 02:19:03 ALARM-401 thermal\n- 2024-08-12 02:19:04 ERROR overheat pack_main\nEXCERPTS:\n- [NOTES.md] Title: Housing bore drift / pack venting | Domain: machining | What went wrong: | The CNC bore is 0.18 mm undersize and the surface is smeared. After assembly, the last three 21700 cells in the 6S2P pack vented during a 2C charge. | Objective:\n- [bore_finish.nc] (REV C bore finish) | T4 M6 | S4200 M3 | G1 Z-12.0 F180 | M30\n- [charger.log] 2024-08-12 02:14:01 INFO charger ready | 2024-08-12 02:16:10 INFO CC 2.0C pack_main | 2024-08-12 02:18:44 WARN cell_5 imbalance 42 mV |", "failure_category": "reagent_material_flaw", "gold_cause": "The reused 6 mm carbide end mill was worn beyond its tool-life limit. Flank wear and minor flute chipping increased cutting forces, vibration, and tool deflection during slot finishing. The worn cutter rubbed instead of shearing cleanly, creating burrs, poor surface finish, and undersized slots in later parts.\n", "gold_hypothesis_index": 4, "group_id": "yaml:RF-009", "id": "RF-009-AUG512", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Tool wear and dulling raise cutting force; undersize/smear often track tool condition and coolant.", "Instrumentation artifacts (calibration, aliasing, grounding) compete with process faults."], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "cnc_milling", "tabular_stats": [43.315, 51.73901904430221, 5.88, 160.0, 18.0, 32.906, 53.5438999700246, 0.01, 160.0, 15.0, 47.08708333333333, 45.925591039488964, 0.18, 160.0, 24.0, 5.587999999999999, 2.0, 1.4142135623730951, 0.0, 4.0, 5.0, 52.160000000000004, 31.96989834203418, 25.1, 94.0, 5.0, 2.02, 1.2140016474453401, 1.02, 3.6, 5.0, 0.0], "title": "CNC End Mill Wear Drives Dimensional Drift and Burr Formation", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Heat dissipation path obstructed / cooling insufficient at low rotational speed", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Process physics consistent with a different AI4I failure mode than the label", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Power overload from high torque at high RPM rather than cooling deficit", "Tool wear failure: elevated torque from dull cutter misread as thermal trip"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process-temperature probe lag under-reporting while cooling is actually adequate", "Process physics consistent with a different AI4I failure mode than the label", "Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Power overload from high torque at high RPM rather than cooling deficit"], "domain": "manufacturing_industrial", "evidence": "Operator dump summary after cleanse: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '302.0', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1294', 'Torque [Nm]': '52.6', 'Tool wear [min]': '26', 'Failure Type': 'Heat Dissipation Failure', 'Machine failure': '1'}. Failure type field=Heat Dissipation Failure.", "failure_category": "reagent_material_flaw", "gold_cause": "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "gold_hypothesis_index": 0, "group_id": "harvest:HV-SYN-0021", "id": "HV-SYN-0021-X1-P1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [302.0, 311.8, 1294.0, 52.6, 26.0, 1.0, 331.23333333333335, 1.0, 1294.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 21", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: M", "Lubrication starvation after interval miss"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '302.2', 'Process temperature [K]': '311.5', 'Rotational speed [rpm]': '1734', 'Torque [Nm]': '25.0', 'Tool wear [min]': '116', 'Machine failure': '0'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0093", "id": "HV-AI4I-0093", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3660.0, 302.2, 311.5, 1734.0, 25.0, 116.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 512.3916666666667, 0.0, 3660.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 93", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: L", "Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.4', 'Process temperature [K]': '311.9', 'Rotational speed [rpm]': '1298', 'Torque [Nm]': '63.9', 'Tool wear [min]': '59', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0060", "id": "HV-AI4I-0060-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4841.0, 303.4, 311.9, 1298.0, 63.9, 59.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 573.2666666666667, 0.0, 4841.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 60", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Actual compressor efficiency degradation", "bad sensor input", "Fuel control problem", "RUL model architecture is invalid", "false prognostics alarm", "The installed compressor discharge pressure sensor was biased low after its harness was disturbed during maintenance. The prognostics model interpreted the biased pressure channel as compressor efficiency loss, producing a false rapid RUL drop.\n"], "difficulty": "hard", "discriminating_followups": ["What observation would rule out: Actual compressor efficiency degradation?", "What observation would rule out: Fuel control problem?"], "distractors": ["Actual compressor efficiency degradation", "Fuel control problem", "RUL model architecture is invalid", "bad sensor input", "false prognostics alarm"], "domain": "mechanical", "evidence": "Messy dump distilled to:\nTITLE: Turbofan Health Monitor Flags False Degradation from Compressor Sensor Bias\nDOMAIN: aerospace / engine_prognostics\nGOAL: Diagnose why a turbofan prognostics model predicts a rapid remaining-useful-life drop during a ground test. The engine should show stable health indicators after a routine borescope inspection.\nPROTOCOL:\n1. Run a standard ground-test profile after maintenance \u2014 Sensors include fan speed, compressor discharge pressure, exhaust gas temperature, and fuel flow.\n2. Feed normalized sensor channels into the RUL estimator \u2014 The estimator was trained on C-MAPSS-like degradation trajectories.\n3. Cross-check suspect channels against redundant gauges \u2014 A portable calibrated pressure transducer is installed after the model flags rapid degradation.\nTELEMETRY_EXTRACT:\n[prognostics_output] Model output and sensor residuals; obs: RUL collapses mainly because compressor pressure appears low.; obs: Temperature and fuel-flow channels do not show a matching degradation pattern.; n=16 mean=27.68 std=54.74 min=-7.2 max=182  | [calibration_check] Pressure-channel cross-check; obs: Installed sensor reads about 7.5% low.; obs: Portable reference matches expected pressure.; n=4 mean=943 std=30.61 min=890 max=962  | [maintenance_notes] Recent maintenance observations; obs: Pressure sensor harness was disconnected during borescope access.; obs: Connector shell shows a bent shield clamp and slight moisture residue.; obs: Borescope found no compressor blade damage.\nCONTEXT_HINTS:\n- The RUL model assumes sensor channels are calibrated and comparable to training data.\n- True compressor degradation should also shift exhaust gas temperature and fuel flow.\n- The warning appeared immediately after maintenance, not gradually over flight cycles.\n- The portable pressure transducer was calibrated last month.", "failure_category": "instrumentation_artifact", "gold_cause": "The installed compressor discharge pressure sensor was biased low after its harness was disturbed during maintenance. The prognostics model interpreted the biased pressure channel as compressor efficiency loss, producing a false rapid RUL drop.\n", "gold_hypothesis_index": 5, "group_id": "yaml:IN-012", "id": "IN-012-AUG797", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Maintenance disturbed the compressor pressure sensor connector", "The installed pressure channel developed a low bias", "Low compressor pressure residual mimicked compressor efficiency degradation", "Tool wear and dulling raise cutting force; undersize/smear often track tool condition and coolant.", "Instrumentation artifacts (calibration, aliasing, grounding) compete with process faults."], "source_file": null, "source_type": "expert_authored_synthetic_paraphrase", "split": "validation", "subdomain": "engine_prognostics", "tabular_stats": [27.68125, 54.73535441958424, -7.2, 182.0, 16.0, 943.0, 30.610455730027933, 890.0, 962.0, 4.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Turbofan Health Monitor Flags False Degradation from Compressor Sensor Bias", "why_alternatives_fail": null}
{"category_label": "reagent_material_flaw", "causal_chain": [], "competing_hypotheses": ["Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Tool wear failure: elevated torque from dull cutter misread as thermal trip", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Process physics consistent with a different AI4I failure mode than the label", "Power overload from high torque at high RPM rather than cooling deficit", "Heat dissipation path obstructed / cooling insufficient at low rotational speed"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Process-temperature probe lag under-reporting while cooling is actually adequate", "Air-temperature sensor bias of +5 K mimicking a true process delta-T rise", "Power overload from high torque at high RPM rather than cooling deficit", "Tool wear failure: elevated torque from dull cutter misread as thermal trip"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial sensor readings extract (AI4I-style). Key readings: {'Air temperature [K]': '298.2', 'Process temperature [K]': '308.3', 'Rotational speed [rpm]': '1345', 'Torque [Nm]': '48.8', 'Tool wear [min]': '74', 'Failure Type': 'Heat Dissipation Failure', 'Machine failure': '1'}. Failure type field=Heat Dissipation Failure.", "failure_category": "reagent_material_flaw", "gold_cause": "Heat dissipation path obstructed / cooling insufficient at low rotational speed", "gold_hypothesis_index": 5, "group_id": "harvest:HV-SYN-0009", "id": "HV-SYN-0009-X1-P1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [298.2, 308.3, 1345.0, 48.8, 74.0, 1.0, 345.8833333333334, 1.0, 1345.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 9", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Machine failure label: M", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss"], "domain": "manufacturing_industrial", "evidence": "Operator dump summary after cleanse: Industrial sensor readings extract (AI4I-style). Key readings: {'Air temperature [K]': '302.0', 'Process temperature [K]': '309.7', 'Rotational speed [rpm]': '1386', 'Torque [Nm]': '62.7', 'Tool wear [min]': '142', 'Machine failure': '1'}. Labeled fault class=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 3, "group_id": "harvest:HV-AI4I-0019", "id": "HV-AI4I-0019-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4354.0, 302.0, 309.7, 1386.0, 62.7, 142.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 546.5333333333333, 0.0, 4354.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 19", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Machine failure label: L", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Thermal expansion bind at operating temperature only", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss"], "domain": "manufacturing_industrial", "evidence": "Post-cleanse anomaly-oriented summary: Industrial sensor readings extract (AI4I-style). Key readings: {'Air temperature [K]': '302.3', 'Process temperature [K]': '311.4', 'Rotational speed [rpm]': '1333', 'Torque [Nm]': '66.7', 'Tool wear [min]': '205', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0053", "id": "HV-AI4I-0053-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3944.0, 302.3, 311.4, 1333.0, 66.7, 205.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0, 513.7833333333333, 0.0, 3944.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 53", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Gauge foil alloy is off-spec", "Unit scaling mismatch between transducer and PLC/engineering units", "50/60 Hz ground-loop / EMI into bridge", "Excitation supply sag only", "Beam modulus wrong in FEA"], "difficulty": "easy", "discriminating_followups": ["Does a battery-powered isolated amp kill the hum?", "Is hum coherent with line frequency on FFT?"], "distractors": ["Gauge foil alloy is off-spec", "Excitation supply sag only", "Beam modulus wrong in FEA", "Unit scaling mismatch between transducer and PLC/engineering units"], "domain": "physics_priors", "evidence": "Strain gauge bridge shows 50/60 Hz hum amplitude rivaling the expected mechanical signal after a new DAQ share of building ground. Static load test on insulated bench is clean.", "failure_category": "instrumentation_artifact", "gold_cause": "50/60 Hz ground-loop / EMI into bridge", "gold_hypothesis_index": 2, "group_id": "prior:PRIOR-014", "id": "PRIOR-014", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["Shared grounds create loop currents that couple into low-level bridges.", "If insulated bench is clean but plant install hums, prefer EMI/grounding over gauge foil chemistry."], "source_file": null, "source_type": "synthetic_physics_prior", "split": "validation", "subdomain": "ground_loop_noise", "tabular_stats": [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Synthetic prior: ground_loop_noise", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Intermittent wiring fault correlated with vibration, not random label noise", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label", "Machine failure label: L"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Shop-floor measurement digest (AI4I-style). Salient channels: {'Air temperature [K]': '302.3', 'Process temperature [K]': '311.4', 'Rotational speed [rpm]': '1333', 'Torque [Nm]': '66.7', 'Tool wear [min]': '205', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 5, "group_id": "harvest:HV-AI4I-0053", "id": "HV-AI4I-0053-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [3944.0, 302.3, 311.4, 1333.0, 66.7, 205.0, 1.0, 0.0, 0.0, 1.0, 1.0, 0.0, 513.7833333333333, 0.0, 3944.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 53", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Power/torque path indicates overload trip condition", "Tool wear beyond process window elevating cutting torque", "Heat dissipation path obstructed at low rotational speed", "Current-transformer phase error inventing apparent power trip", "Overstrain recipe outside qualified envelope without power-limit hit", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Current-transformer phase error inventing apparent power trip", "Tool wear beyond process window elevating cutting torque", "Overstrain recipe outside qualified envelope without power-limit hit", "Heat dissipation path obstructed at low rotational speed"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '298.1', 'Process temperature [K]': '308.7', 'Rotational speed [rpm]': '1629', 'Torque [Nm]': '62.6', 'Tool wear [min]': '47', 'Failure Type': 'Power Failure', 'Machine failure': '1'}. Labeled fault class=Power Failure.", "failure_category": "instrumentation_artifact", "gold_cause": "Power/torque path indicates overload trip condition", "gold_hypothesis_index": 0, "group_id": "harvest:HV-SYN-0003", "id": "HV-SYN-0003-X1-P1", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [298.1, 308.7, 1629.0, 62.6, 47.0, 1.0, 391.06666666666666, 1.0, 1629.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 3", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Heat dissipation path obstructed at low rotational speed", "Current-transformer phase error inventing apparent power trip", "Process physics consistent with a different AI4I failure mode than the label", "Power/torque path indicates overload trip condition", "Tool wear beyond process window elevating cutting torque", "Overstrain recipe outside qualified envelope without power-limit hit"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Tool wear beyond process window elevating cutting torque", "Current-transformer phase error inventing apparent power trip", "Heat dissipation path obstructed at low rotational speed", "Overstrain recipe outside qualified envelope without power-limit hit", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Cleaned extract (paraphrase): Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '298.1', 'Process temperature [K]': '308.7', 'Rotational speed [rpm]': '1629', 'Torque [Nm]': '62.6', 'Tool wear [min]': '47', 'Failure Type': 'Power Failure', 'Machine failure': '1'}. Failure type field=Power Failure.", "failure_category": "instrumentation_artifact", "gold_cause": "Power/torque path indicates overload trip condition", "gold_hypothesis_index": 3, "group_id": "harvest:HV-SYN-0003", "id": "HV-SYN-0003-X1-P2", "intervention": null, "license": "CC-BY-4.0", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "synthetic_schema_harvest", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [298.1, 308.7, 1629.0, 62.6, 47.0, 1.0, 391.06666666666666, 1.0, 1629.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 3", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Machine failure label: M", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only"], "domain": "manufacturing_industrial", "evidence": "Evidence packet distilled from messy sensors: Shop-floor measurement digest (AI4I-style). Salient channels: {'Air temperature [K]': '302.4', 'Process temperature [K]': '310.3', 'Rotational speed [rpm]': '1378', 'Torque [Nm]': '54.3', 'Tool wear [min]': '165', 'Machine failure': '1'}. Failure type field=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 2, "group_id": "harvest:HV-AI4I-0015", "id": "HV-AI4I-0015-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4450.0, 302.4, 310.3, 1378.0, 54.3, 165.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 555.1666666666666, 0.0, 4450.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 15", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Intermittent wiring fault correlated with vibration, not random label noise", "Machine failure label: L", "Process physics consistent with a different AI4I failure mode than the label"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Process physics consistent with a different AI4I failure mode than the label"], "domain": "manufacturing_industrial", "evidence": "Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '303.4', 'Process temperature [K]': '311.9', 'Rotational speed [rpm]': '1298', 'Torque [Nm]': '63.9', 'Tool wear [min]': '59', 'Machine failure': '1'}. Failure type field=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0060", "id": "HV-AI4I-0060", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4841.0, 303.4, 311.9, 1298.0, 63.9, 59.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 573.2666666666667, 0.0, 4841.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 60", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise", "Machine failure label: M", "Thermal expansion bind at operating temperature only"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold", "Thermal expansion bind at operating temperature only"], "domain": "manufacturing_industrial", "evidence": "Operator dump summary after cleanse: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '303.4', 'Process temperature [K]': '311.8', 'Rotational speed [rpm]': '1316', 'Torque [Nm]': '50.9', 'Tool wear [min]': '69', 'Machine failure': '1'}. Labeled fault class=M.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: M", "gold_hypothesis_index": 4, "group_id": "harvest:HV-AI4I-0041", "id": "HV-AI4I-0041-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4845.0, 303.4, 311.8, 1316.0, 50.9, 69.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 574.8416666666667, 0.0, 4845.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 41", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Thermal expansion bind at operating temperature only", "Lubrication starvation after interval miss", "Bearing defect progressing below vibration alarm threshold", "Machine failure label: L"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Process physics consistent with a different AI4I failure mode than the label", "Thermal expansion bind at operating temperature only", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss", "Intermittent wiring fault correlated with vibration, not random label noise"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Industrial telemetry extract (AI4I-style). Key readings: {'Air temperature [K]': '301.8', 'Process temperature [K]': '310.2', 'Rotational speed [rpm]': '2825', 'Torque [Nm]': '5.8', 'Tool wear [min]': '215', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 5, "group_id": "harvest:HV-AI4I-0002", "id": "HV-AI4I-0002-P2", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [4297.0, 301.8, 310.2, 2825.0, 5.8, 215.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 663.0666666666667, 0.0, 4297.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 2", "why_alternatives_fail": null}
{"category_label": "instrumentation_artifact", "causal_chain": [], "competing_hypotheses": ["Thermal expansion bind at operating temperature only", "Machine failure label: L", "Intermittent wiring fault correlated with vibration, not random label noise", "Process physics consistent with a different AI4I failure mode than the label", "Bearing defect progressing below vibration alarm threshold", "Lubrication starvation after interval miss"], "difficulty": "easy", "discriminating_followups": ["Was rotational speed below the cooling design point when the fault flagged?", "What was tool-wear minutes relative to the qualified tool-life limit?", "Did torque\u00d7speed approach the rated power envelope?"], "distractors": ["Lubrication starvation after interval miss", "Thermal expansion bind at operating temperature only", "Process physics consistent with a different AI4I failure mode than the label", "Intermittent wiring fault correlated with vibration, not random label noise", "Bearing defect progressing below vibration alarm threshold"], "domain": "manufacturing_industrial", "evidence": "Compact multimodal extract \u2014 stats only: Industrial sensor readings extract (AI4I-style). Salient channels: {'Air temperature [K]': '295.7', 'Process temperature [K]': '306.2', 'Rotational speed [rpm]': '2270', 'Torque [Nm]': '14.6', 'Tool wear [min]': '149', 'Machine failure': '1'}. Labeled fault class=L.", "failure_category": "instrumentation_artifact", "gold_cause": "Machine failure label: L", "gold_hypothesis_index": 1, "group_id": "harvest:HV-AI4I-0031", "id": "HV-AI4I-0031-P1", "intervention": null, "license": "CC-BY-4.0 (AI4I source data: CC BY 4.0, UCI)", "packet_id": null, "physics_priors": ["At low RPM, convective cooling drops \u2014 heat dissipation failures become more likely if process-air delta-T is high.", "Tool wear increases cutting force / torque for the same feed \u2014 overstrain risk rises.", "Power \u2248 torque \u00d7 angular speed; trips near rated power suggest overload not sensor ghosting alone."], "source_file": null, "source_type": "harvested_public_ai4i", "split": "validation", "subdomain": "predictive_maintenance", "tabular_stats": [904.0, 295.7, 306.2, 2270.0, 14.6, 149.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 328.4583333333333, 0.0, 2270.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], "title": "Harvested industrial failure sample 31", "why_alternatives_fail": null}