Llama-3.1-8B-Instruct finetuned on paraphrase-augmented TOFU, retain/reference models, and checkpoints unlearned with open-unlearning methods.
-
sungjuncho/TOFU-para
Updated • 100 -
sungjuncho/Llama-3.1-8B-Instruct-TOFU-para-full-seed0
Text Generation • 8B • Updated • 447 -
sungjuncho/Llama-3.1-8B-Instruct-TOFU-para-x2-full-seed0
Text Generation • 8B • Updated • 458 -
sungjuncho/Llama-3.1-8B-Instruct-TOFU-para-x2-retain90-seed0
Text Generation • 8B • Updated • 460