Lexsi/audit-harden-SaLoRAHardenTrainer-llama32-3b-legal
3B • Updated
AI & ML interests
Frontier research around Safe and aligned intelligence
Recent Activity
View all activity
Papers
Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution
$C$-$ΔΘ$: Circuit-Restricted Weight Arithmetic for Selective Refusal
Lexsi/audit-harden-TARModel-llama32-3b-legal
3B • Updated • 15
Lexsi/audit-harden-DOORTrainer-llama32-3b-legal
3B • Updated
Lexsi/audit-harden-DeRTaTrainer-llama32-3b-legal
3B • Updated
Lexsi/audit-harden-SurgeryTrainer-llama32-3b-legal
3B • Updated
Lexsi/audit-harden-AsFTTrainer-llama32-3b-legal
3B • Updated
Lexsi/audit-harden-undefended-SFT-llama32-3b-legal
3B • Updated • 16
Lexsi/audit-recover-apply_prepost_merge-llama32-3b-legal
3B • Updated • 11
Lexsi/audit-recover-apply_nlsr-llama32-3b-legal
3B • Updated • 14
Lexsi/audit-recover-apply_lssf-llama32-3b-legal
3B • Updated • 37
Lexsi/audit-recover-somf_merge-llama32-3b-legal
3B • Updated • 11
Lexsi/audit-recover-apply_safemerge-llama32-3b-legal
3B • Updated • 15
Lexsi/audit-recover-apply_safe_lora-llama32-3b-legal
3B • Updated • 53
Lexsi/audit-harden-undefended-SFT-qwen3-4b-dolly
4B • Updated • 30
Lexsi/audit-harden-SaLoRAHardenTrainer-llama31-8b-code
8B • Updated • 16
Lexsi/audit-harden-SaLoRAHardenTrainer-llama32-3b-gsm8k
3B • Updated • 16
Lexsi/qwen3-4b-dolly-dpo-recover
4B • Updated • 8
Lexsi/audit-recover-apply_lssf-llama31-8b-code
8B • Updated • 41
Lexsi/audit-harden-undefended-SFT-llama31-8b-medical
8B • Updated • 31
Lexsi/audit-recover-apply_lssf-llama32-3b-gsm8k
3B • Updated • 22
Lexsi/audit-harden-undefended-SFT-llama31-8b-dolly
8B • Updated • 29
Lexsi/audit-harden-undefended-SFT-llama32-3b-medical
3B • Updated • 31
Lexsi/qwen3-4b-code-dpo-recover
4B • Updated • 16
Lexsi/audit-harden-undefended-SFT-llama32-3b-code
3B • Updated • 30
Lexsi/audit-recover-apply_safe_lora-llama31-8b-medical
8B • Updated • 22
Lexsi/audit-harden-undefended-SFT-llama32-3b-dolly
3B • Updated • 34
Lexsi/audit-recover-apply_safe_lora-llama31-8b-dolly
8B • Updated • 9
Lexsi/audit-recover-apply_safe_lora-llama32-3b-medical
3B • Updated • 16
Lexsi/audit-recover-apply_safe_lora-llama32-3b-code
3B • Updated • 42
Lexsi/audit-recover-apply_safe_lora-llama32-3b-dolly
3B • Updated • 32