Tootfinder

Opt-in global Mastodon full text search. Join the index!

No exact results. Similar results found.
@compfu@mograph.social
2025-06-24 20:39:37

I wonder how this can be leveraged in our #VFX industry.
Can you position a company as the guys who’ll fix stuff that another team messed up because they thought they can do an effects shot with Midjourney but then ended up with nothing to show once the director wanted a simple revision like turning that object a tiny bit more or making the cool explosion more „swooosh“ instead of „wooosh“?

…

Security Writer (@SecurityWriter@infosec.exchange)
Would you like another prediction? Generative AI as we know it is a bubble. It’s unsustainable. Relying on it now will result in very expensive work to undo that damage in years to come. And we’re going to live with the accuracy debt for decades. We are going to have an enormous skills and critical thinking shortage in the very near future, once the layoffs have slowed, and organisations start to cave in under the weight of their own shortsightedness. Consultancy demand will go through…

@arXiv_csMA_bot@mastoxiv.page
2025-06-12 07:49:11

Effective Red-Teaming of Policy-Adherent Agents
Itay Nakash, George Kour, Koren Lazar, Matan Vetzler, Guy Uziel, Ateret Anaby-Tavor
arxiv.org/abs/2506.09600

Effective Red-Teaming of Policy-Adherent Agents
Task-oriented LLM-based agents are increasingly used in domains with strict policies, such as refund eligibility or cancellation rules. The challenge lies in ensuring that the agent consistently adheres to these rules and policies, appropriately refusing any request that would violate them, while still maintaining a helpful and natural interaction. This calls for the development of tailored design and evaluation methodologies to ensure agent resilience against malicious user behavior. We propos…