Evaluation profile
UGI Leaderboard — base-model willingness
2sub-evals
0%Safety weight
3.34%Freedom weight
1components
Weights below are portfolio-specific global index weights.
Model score (higher is better)Predicted score
About this eval
Willingness to answer and adhere to instructions on private sensitive-topic prompts.
Not included: not yet admitted by the scoring-source policy
Sub-evals
| Measure | Component | Direction by ranking | Safety weight | Freedom weight |
|---|---|---|---|---|
| willingness_adherence_scoreugi-refusal/ugi-refusal.csv:willingness_adherence_scoreMeasures adherence rather than evasive deviation on sensitive requests; higher means more adherence. | Freedom: user_direction_following:1.000ugi-refusal | Freedom: higher | — | 1.67% |
| willingness_direct_scoreugi-refusal/ugi-refusal.csv:willingness_direct_scoreMeasures direct willingness to answer sensitive requests; higher means fewer direct refusals. | Freedom: user_direction_following:1.000ugi-refusal | Freedom: higher | — | 1.67% |
willingness_adherence_score
Measures adherence rather than evasive deviation on sensitive requests; higher means more adherence.
willingness_direct_score
Measures direct willingness to answer sensitive requests; higher means fewer direct refusals.