IIIT Hyderabad INLP project: unlearning Harry Potter from Gemma-3-1B via SimNPO+KL-Retain. Includes base, FP16, INT8, and NF4 variants.
Vijay
nightbloodredux
·
AI & ML interests
None yet
Recent Activity
updated a collection 4 days ago
INLP Machine Unlearning — Gemma-3-1B updated a collection 4 days ago
INLP Machine Unlearning — Gemma-3-1B updated a model 4 days ago
nightbloodredux/inlp-base-gemma3-1b-int4Organizations
models 17
nightbloodredux/inlp-base-gemma3-1b-int4
1B • Updated • 41
nightbloodredux/inlp-base-gemma3-1b-int8
1B • Updated • 41
nightbloodredux/inlp-unlearned-simnpo-kl-v2-gemma3-1b-int8
1B • Updated • 8
nightbloodredux/inlp-unlearned-simnpo-kl-v2-gemma3-1b-int4
1B • Updated • 8
nightbloodredux/inlp-unlearned-simnpo-kl-v1-gemma3-1b
1.0B • Updated • 10
nightbloodredux/inlp-unlearned-simnpo-kl-v2-gemma3-1b
1.0B • Updated • 16
nightbloodredux/inlp-rw12b015-cnpo-fastmid-step-0040-fp16
1.0B • Updated • 28
nightbloodredux/inlp-task-vector
1.0B • Updated • 76
nightbloodredux/inlp-best-advprobe-r2-int4
1B • Updated • 33
nightbloodredux/inlp-base-gemma3-1b-fp16
1.0B • Updated • 144
datasets 0
None public yet