Tootfinder

Concept activation vectors: a unifying view and adversarial attacks
Ekkehard Schnoor, Malik Tiomoko, Jawher Said, Alex Jung, Wojciech Samek
https://arxiv.org/abs/2509.22755 http…

Concept activation vectors: a unifying view and adversarial attacks
Concept Activation Vectors (CAVs) are a tool from explainable AI, offering a promising approach for understanding how human-understandable concepts are encoded in a model's latent spaces. They are computed from hidden-layer activations of inputs belonging either to a concept class or to non-concept examples. Adopting a probabilistic perspective, the distribution of the (non-)concept inputs induces a distribution over the CAV, making it a random vector in the latent space. This enables us to der…

Closing the Safety Gap: Surgical Concept Erasure in Visual Autoregressive Models
Xinhao Zhong, Yimin Zhou, Zhiqi Zhang, Junhao Li, Yi Sun, Bin Chen, Shu-Tao Xia, Ke Xu
https://arxiv.org/abs/2509.22400 …

Closing the Safety Gap: Surgical Concept Erasure in Visual Autoregressive Models
The rapid progress of visual autoregressive (VAR) models has brought new opportunities for text-to-image generation, but also heightened safety concerns. Existing concept erasure techniques, primarily designed for diffusion models, fail to generalize to VARs due to their next-scale token prediction paradigm. In this paper, we first propose a novel VAR Erasure framework VARE that enables stable concept erasure in VAR models by leveraging auxiliary visual tokens to reduce fine-tuning intensity. B…

No, SNAP Benefits Aren’t Mostly Used by Immigrants | WIRED #SNAP

No, SNAP Benefits Aren’t Mostly Used by Immigrants
SNAP benefits are set to run out on Saturday. Far-right influencers and extremists are incorrectly claiming that immigrants are the main recipients of food stamps.

Schur log-concavity and the quantum Pascal triangle
\'Alvaro Guti\'errez, Christian Krattenthaler
https://arxiv.org/abs/2509.22648 https://arxiv.or…

Schur log-concavity and the quantum Pascal triangle
We say a sequence $f_0, f_1, f_2, \ldots$ of symmetric functions is Schur log-concave if $f_n^2 - f_{n-1}f_{n+1}$ is Schur positive for all $n\ge1$. We conjecture that a very general class of sequences of Schur functions satisfies this property, and show it for sequences of Schur functions indexed by partitions with growing first part and column. Our findings are related to work of Lam, Postnikov and Pylyavskyy on Schur positivity, and of Butler, Sagan, and the second author on $q$-log-concavit…

Aerodynamic performance and robustness of a nature-inspired concept for a micro-scale wind turbine
J. M. Catal\'an, G. Arranz, M. Moriche, M. Guerrero-Hurtado, M. Garc\'ia-Villalba, O. Flores
https://arxiv.org/abs/2509.24998

Aerodynamic performance and robustness of a nature-inspired concept for a micro-scale wind turbine
We present direct numerical simulations of a novel concept for a micro-scale wind turbine, inspired in the mechanics of the auto-rotation of winged seeds. In this nature-inspired concept the turbine blades have two degrees of freedom: the pitch and the elevation (or coning) angles. These allow the blade to vary its attitude with respect to the incoming velocity seen by the blade (i.e., the tip-speed ratio, $λ$). In order to validate this new concept, we perform numerical simulations of the cou…

Real conspiracies tend to come out, but some of them take a while. Information on the Iran/Contra scandal broke out about 5 years after the conspiracy started. That would have taken several hundred people to carry out, so it was somewhat hard to hide. Even so, they largely got away with it.
The moon landing conspiracy theory would have taken thousands of people, so it would have come out more quickly. Since we have an example of a real secret program of a similar scale as what would be required to fake a moon landing (that is, the Manhattan project), we know that the fake moon landing conspiracy theory is not true. (There's also the literally tons of evidence in the form of rocks and other samples, and all kinds of other ways to debunk the claim.)
Could Kash Patel's FBI have been trying really hard to entrap people into carrying out terrorist attacks in order to justify #Trump's occupation of DC? Could they have helped a guy plan an attack then just failed to arrest him? There are reasonable scenarios that fall in between malice and incompetence while still indicating some level of false flag.
Could someone have just snapped and ambushed some guardsmen without any involvement from the FBI? Yeah, totally. The US is a country full of guns with a completely non-functional mental health system. Someone coming from a country that the US destroyed, twice, could have a lot of untreated trauma. Might they see the national guard as a threat (even if that wasn't totally true)? Yeah, they were deployed to threaten people (even when they were just picking up trash). The point was to incite this kind of response. It's completely reasonable to believe that the FBI would not need to be involved at all, that this would just be the stochastic response they were looking for.
So the point here is that everything is on the table, nothing is really known, nothing should be surprising, and no matter what it's Trump's fault. This is exactly the escalation he was looking for. If he didn't get it naturally, he would also have had ways of making it happen.
He will use this in exactly the same way as the Reichstag fire, to drive a wedge between liberals and radicals. Don't fall for it.
Edit:
There are plausible reasons to not believe the official narrative at all right now, or maybe ever. The official narrative is also plausible, but there are plausible reasons to disagree with the response even if the official story is true. It is unnecessary to resort to conspiracy thinking in order to account for what happened and to disagree with the response. But it is also understandable why someone might jump immediately to a conspiracy given the circumstances.

Shabana Mahmood, the latest in the line of migrant heritage Home Secretaries to show just how tough she can be on.... migrants.
Simultaneously playing with fire (we know how this ends) and pulling up the ladder after her (meanwhile basking in privilege).
It would be better if everyone, migrants and natives had to do a bit of community service.

Shabana Mahmood to demand migrants earn right to settled status
New tests will include learning English to a high standard, paying National Insurance and not claiming benefits.

Privy: Envisioning and Mitigating Privacy Risks for Consumer-facing AI Product Concepts
Hao-Ping Lee, Yu-Ju Yang, Matthew Bilik, Isadora Krsek, Thomas Serban von Davier, Kyzyl Monteiro, Jason Lin, Shivani Agarwal, Jodi Forlizzi, Sauvik Das
https://arxiv.org/abs/2509.23525

Privy: Envisioning and Mitigating Privacy Risks for Consumer-facing AI Product Concepts
AI creates and exacerbates privacy risks, yet practitioners lack effective resources to identify and mitigate these risks. We present Privy, a tool that guides practitioners through structured privacy impact assessments to: (i) identify relevant risks in novel AI product concepts, and (ii) propose appropriate mitigations. Privy was shaped by a formative study with 11 practitioners, which informed two versions -- one LLM-powered, the other template-based. We evaluated these two versions of Privy…

There's some kind of right wing conspiracy theory involving the beds from, I assume, the movie Elysium (2013)?
https://www.theverge.com/news/787042/trump-posts-then-pulls-bizarre-ai-video-promoting-medbed-conspiracy

Trump posts, then pulls bizarre AI video promoting MedBed conspiracy
Trump shared an AI-generated video in which he announced a MedBed hospital and national MedBed card.

Diamonds in the rough: Transforming SPARCs of imagination into a game concept by leveraging medium sized LLMs
Julian Geheeb, Farhan Abid Ivan, Daniel Dyrda, Miriam Ansch\"utz, Georg Groh
https://arxiv.org/abs/2509.24730

Diamonds in the rough: Transforming SPARCs of imagination into a game concept by leveraging medium sized LLMs
Recent research has demonstrated that large language models (LLMs) can support experts across various domains, including game design. In this study, we examine the utility of medium-sized LLMs, models that operate on consumer-grade hardware typically available in small studios or home environments. We began by identifying ten key aspects that contribute to a strong game concept and used ChatGPT to generate thirty sample game ideas. Three medium-sized LLMs, LLaMA 3.1, Qwen 2.5, and DeepSeek-R1, …