Abliteration Leaderboard
Capability Preservation
Tool Compatibility
Community Models
v1 vs v2
About
Abliteration Leaderboard
Capability Preservation
Tool Compatibility
Community Models
v1 vs v2
About
Minimum ASR (%)
↺
0
100
Maximum KL Divergence
↺
0
2
ASR by Model
KL Divergence vs Refusals