uid stringclasses 10
values | dataset_id stringclasses 1
value | region stringclasses 9
values | training_ramp_signal stringclasses 3
values | grid_capacity_signal stringclasses 3
values | water_cooling_capacity_signal stringclasses 3
values | cost_spike_signal stringclasses 3
values | ops_summary stringclasses 10
values | question stringclasses 1
value | ground_truth_label stringclasses 2
values | ground_truth_rationale stringclasses 10
values | outcome_signal stringclasses 6
values | source_citation stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
AEW-001 | ai-training-energy-water-coherence-risk-v0.1 | US-West | stable | high | high | low | Training steady with ample power and water. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | coherent | Capacity supports demand. | stable | Synthetic |
AEW-002 | ai-training-energy-water-coherence-risk-v0.1 | US-West | spiking | low | low | high | Training ramp with constrained grid and cooling water. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | incoherent | Demand exceeds supply. | shutdown_risk | Synthetic |
AEW-003 | ai-training-energy-water-coherence-risk-v0.1 | EU-North | rising | moderate | high | moderate | Power moderate and water strong. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | coherent | Manageable ramp. | stable | Synthetic |
AEW-004 | ai-training-energy-water-coherence-risk-v0.1 | India | spiking | low | moderate | high | Rapid training expansion strains grid. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | incoherent | Energy constraint. | cost_spike | Synthetic |
AEW-005 | ai-training-energy-water-coherence-risk-v0.1 | Canada | stable | high | high | low | Cold region with strong cooling. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | coherent | Aligned. | stable | Synthetic |
AEW-006 | ai-training-energy-water-coherence-risk-v0.1 | US-South | spiking | moderate | low | high | Water cooling limited. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | incoherent | Cooling bottleneck. | thermal_limit | Synthetic |
AEW-007 | ai-training-energy-water-coherence-risk-v0.1 | EU-West | stable | moderate | moderate | low | Balanced. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | coherent | No mismatch. | stable | Synthetic |
AEW-008 | ai-training-energy-water-coherence-risk-v0.1 | China | spiking | low | low | high | Large ramp with constrained utilities. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | incoherent | Severe mismatch. | curtailment | Synthetic |
AEW-009 | ai-training-energy-water-coherence-risk-v0.1 | US-Northwest | rising | high | high | low | Hydro and cooling strong. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | coherent | Resilient. | stable | Synthetic |
AEW-010 | ai-training-energy-water-coherence-risk-v0.1 | MiddleEast | spiking | high | low | moderate | Power ample but water scarce. | Is training demand vs energy and water coherent. Answer coherent or incoherent. | incoherent | Cooling risk. | efficiency_drop | Synthetic |
What this repo is for
Detect when AI training expansion outpaces
power supply or cooling water
before cost spikes or forced curtailment.
- Downloads last month
- 19