metadata
license: apache-2.0
language: en
tags:
- agentic-ai
- evaluation
- scoring
- regression
- quality
pretty_name: AIF Evaluations
size_categories: n<1K
AIF Evaluations
14-dimension agent evaluation datasets with historical trends, regression detection, and quality scoring.
Evaluation Dimensions
| # | Dimension | Description |
|---|---|---|
| 1 | Accuracy | Task completion correctness |
| 2 | Efficiency | Token usage and time optimization |
| 3 | Safety | Content filtering and constraint adherence |
| 4 | Reliability | Consistency across repeated runs |
| 5 | Adaptability | Performance on novel tasks |
| 6 | Explainability | Decision transparency |
| 7 | Robustness | Graceful error handling |
| 8 | Latency | Response time compliance |
| 9 | Cost | Budget adherence |
| 10 | Compliance | Policy and governance adherence |
| 11 | Collaboration | Multi-agent coordination quality |
| 12 | Creativity | Novel solution generation |
| 13 | Memory | Context retention and recall |
| 14 | Learning | Improvement over time |
Usage
from datasets import load_dataset
ds = load_dataset("FrostyJay7813/aif-evaluations")