A fully uncensored Hermes 3 with lorablation! The lorablated model directly answers questions without any tweaking. Here's how it was made: 1/ Create a LoRA adapter based on Llama 3.1 8B Instruct and my abliterated version. 2/ Apply it to Hermes 3 from @NousResearch using

→ View original post on X — @maximelabonne,