huihui-ai/Llama-3.3-70B-Instruct-abliterated

huihui-ai/Llama-3.3-70B-Instruct-abliterated is a 70 billion parameter instruction-tuned causal language model derived from Meta's Llama-3.3-70B-Instruct, featuring a 32768 token context length. This model has been specifically modified using an "abliteration" technique to remove refusal behaviors, aiming to provide uncensored responses. It is primarily designed for use cases requiring direct answers without built-in content moderation or refusal mechanisms.