Large language models (LLMs) require iterative updates to address the outdated information problem, where LLM unlearning offers an approach for selective removal. However, mainstream unlearning methods primarily rely on fine-tuning techniques, which often lack precision in targeted unlearning and struggle to balance unlearning efficacy with general ability under massive and sequential settings. To bridge this gap, in this work, we introduce UniErase, a novel unlearning framework that demonstrates precision and balanced performances between knowledge unlearning and ability retaining. We first propose the Unlearning Token, which is optimized to steer LLMs toward a forgetting space. To achieve concrete unlearning behaviors, we further introduce the lightweight Unlearning Edit to efficiently associate the unlearning targets with this meta-token. Serving as a new unlearning paradigm via editing, UniErase achieves outstanding performances across batch, sequential, and precise unlearning tasks under fictitious and real-world knowledge scenarios. On the TOFU benchmark, compared with 8 baselines, UniErase, modifying only $\sim$ \textbf{3.66%} of the LLM parameters, outperforms the previous best-forgetting baseline by \textbf{$\sim$ 4.01$\times$} for \textbf{model ability} with even higher unlearning efficacy. Similarly, UniErase, with better ability retention, also surpasses the previous best-retaining method by \textbf{35.96%} for \textbf{unlearning efficacy}, showing balanced and dual top-tier performances in the current unlearning community.

Manila, Philippines – Leila de Lima was released from detention last month into what the former Philippines senator calls "a whole new world". In 2016, then-President Rodrigo Duterte promised to "destroy" de Lima, one of the loudest critics of his deadly drug war. The president's supporters began targeting the first-term senator and former human rights commissioner – ridiculing her for an alleged romantic affair with her driver, and accusing her of involvement in drug trafficking. In February 2017, she was arrested on drug charges she denies and that international observers have said are politically motivated. "I had this deep sense of disbelief," de Lima told Al Jazeera. "I never thought that Mr Duterte would go to that extent, that length, of jailing me. I thought it would just be daily vilification, personal attacks, attacks against my womanhood."

Is your company using AI? Considering it? A Google search for "artificial empathy" (AE) has over 8 million results. Wiki defines AE as, "the development of AI systems such as companion robots that are able to detect and respond to human emotions." As an executive business professional, I think it's a fascinating concept, making technology more human by teaching it about feelings. However, as I'm also a practicing clinician as a licensed mental health therapist, the struggle comes when I put my therapist hat on and think about the deep, immersive, complicated world of human emotions.