Question 1

How does min-max normalization work?

Accepted Answer

For each benchmark, the minimum raw score among entered models becomes 0, the maximum becomes 1. All scores are scaled linearly between 0 and 1.

Question 2

Can I adjust benchmark weights?

Accepted Answer

Yes, use the sliders below the grid. Weights are normalized to sum to 1.0 automatically. Default equal weights (0.125 each).

Question 3

How is the composite score calculated?

Accepted Answer

Composite = sum over benchmarks of (weight_i × normalized_score_i). Models are ranked by composite score descending.

🧪 LLM Benchmark Score Normalizer

📊 Data Entry