Models That Know How Evaluations Are Designed Score Safer
Paper • 2605.28591 • Published • 6
🕵️🛡️ Evaluation Meta Knowledge
2026 arXiv preprint. Models fine-tuned on documents describing typical evaluation traits show safer behavior by having increased refusal rates and low
-
-
compass-group-tue/sdf_evaluation_traits
Viewer • Updated • 74.8k • 94 • 1 -
compass-group-tue/nemotron-traits
Text Generation • Updated • 38 -
compass-group-tue/Nemotron-Traits-Seed-44
Text Generation • Updated • 11 -
compass-group-tue/Nemotron-Traits-Seed-43
Text Generation • Updated • 28 -
compass-group-tue/nemotron-6-traits
Text Generation • Updated • 19