Skip to content

Benchmark

PashtoBench safety refusal consistency delta (English minus Pashto)

Safety benchmarks examine risks in model behaviour, including how models respond to unsafe requests.

Models measured
4
Scores recorded
4
Sources
1
IQ weight
No weight published

Score distribution

0.289
Mean
0.545
Median
0.395
Std dev
-0.182
10th percentile
0.792
90th percentile
-0.182
Min
-0.1820.792
The band covers the middle 80% of scores. The line marks the median.

Results

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

ModelScore
Claude Sonnet 5column: consistency delta· PashtoBench v1.0 data docs0
GPT 4ocolumn: consistency delta· PashtoBench v1.0 data docs-0.182
Llama 3.3 70B Instructcolumn: consistency delta· PashtoBench v1.0 data docs0.545
Qwen2.5 72B Instructcolumn: consistency delta· PashtoBench v1.0 data docs0.792

Data credits: PashtoBench v1.0, Jawad Noori. Public results/scores numbers; code MIT and data/ CC BY 4.0 grants do not name results/, so no results licence is asserted.