Tootfinder

Hatevolution: What Static Benchmarks Don't Tell Us
Chiara Di Bonaventura, Barbara McGillivray, Yulan He, Albert Mero\~no-Pe\~nuela
https://arxiv.org/abs/2506.12148

Hatevolution: What Static Benchmarks Don't Tell Us
Language changes over time, including in the hate speech domain, which evolves quickly following social dynamics and cultural shifts. While NLP research has investigated the impact of language evolution on model training and has proposed several solutions for it, its impact on model benchmarking remains under-explored. Yet, hate speech benchmarks play a crucial role to ensure model safety. In this paper, we empirically evaluate the robustness of 20 language models across two evolving hate speec…

Congratulations to Professor Braxton Shelley of YDS and the ISM! He has been appointed to the newly endowed George Washington Williams Chair, named for the first African American to graduate from Newton Theological Institution (which continues today as part of Andover Newton Seminary at YDS).
Read about Prof. Shelley's appointment and other YDS faculty accomplishments here.

"Zo neemt het gebruik van groene stroom minder snel toe dan verwacht, schrijft demissionair minister Hermans van Groene Groei en Klimaat (VVD) in een Kamerbrief.
Voornaamste oorzaak is dat de verduurzaming van de industrie moeizaam verloopt en fabrieken dus langer gebruik moeten maken van fossiele brandstoffen als olie, aardgas en steenkool."
♻️ Perpetuum mobile of drogredenering?

Kabinet snoeit flink in plannen voor windenergie op de Noordzee
Na veel slecht nieuws over windenergie op zee stelt het demissionaire kabinet het streefdoel officieel naar beneden bij.

{purrr} has some lesser known functions that make handling of failing function calls easier: safely, quietly, possibly: #rstats…

Wrap a function to capture errors — safely
Creates a modified version of .f that always succeeds. It returns a list with components result and error. If the function succeeds, result contains the returned value and error is NULL. If an error occurred, error is an error object and result is either NULL or otherwise.

It seems like, again, just following the plain logic of the Constitution and Declaration of Independence (which, again, I do not subscribe to), that every law passed under Trump, every supreme court justice appointment by Trump, every supreme court ruling by Trump appointed justices, all the illegal firing, etc, must all, necessarily, be null and void.
And if not following from the insurrection act, or from the oath of office, then following from the Declaration of Independence itself. The logic here being that a constitution is a contract between the people and their government, which the later upholds in order to maintain its legal status. The violation of said laws by the government violates "consent of the governed" (which, again, I have issues with the concept entirely but we're just going to ignore that) and therefore nullifies the authority of that government, granting " the right of the people to alter or to abolish it."
That seems a lot like the hard reset some folks have been looking for. Given that existing flaws allowed this state to be reached, it would also be necessary for the true authority to correct those mistakes before assuming authority that derives from these principles.
Now, personally, I don't subscribe to any of this logic but it's interesting to explore, as an outsider, where the logic goes.

Filter Equivariant Functions: A symmetric account of length-general extrapolation on lists
Owen Lewis, Neil Ghani, Andrew Dudzik, Christos Perivolaropoulos, Razvan Pascanu, Petar Veli\v{c}kovi\'c
https://arxiv.org/abs/2507.08796

Filter Equivariant Functions: A symmetric account of length-general extrapolation on lists
What should a function that extrapolates beyond known input/output examples look like? This is a tricky question to answer in general, as any function matching the outputs on those examples can in principle be a correct extrapolant. We argue that a "good" extrapolant should follow certain kinds of rules, and here we study a particularly appealing criterion for rule-following in list functions: that the function should behave predictably even when certain elements are removed. In functional prog…

Trump is working on regime change in Europe — fact, not conspiracy theory - @…
https://euobserver.com/eu-and-the-world/ar0ec480cc

Null geodesics, causal structure, and matter accretion in Lorentzian-Euclidean black holes
Salvatore Capozziello, Emmanuele Battista, Silvia De Bianchi
https://arxiv.org/abs/2507.08431

Null geodesics, causal structure, and matter accretion in Lorentzian-Euclidean black holes
Recently, we introduced the Lorentzian-Euclidean black hole, a static and spherically symmetric solution of vacuum Einstein equations that exhibits a change in metric signature across the event horizon. In this framework, the analysis of radial trajectories of freely falling bodies proves that the central singularity can be avoided via a mechanism we interpret as atemporality, which is responsible for the shift of the time variable from real to imaginary values. In this paper, we further expl…

PersonaLens: A Benchmark for Personalization Evaluation in Conversational AI Assistants
Zheng Zhao, Clara Vania, Subhradeep Kayal, Naila Khan, Shay B. Cohen, Emine Yilmaz
https://arxiv.org/abs/2506.09902

PersonaLens: A Benchmark for Personalization Evaluation in Conversational AI Assistants
Large language models (LLMs) have advanced conversational AI assistants. However, systematically evaluating how well these assistants apply personalization--adapting to individual user preferences while completing tasks--remains challenging. Existing personalization benchmarks focus on chit-chat, non-conversational tasks, or narrow domains, failing to capture the complexities of personalized task-oriented assistance. To address this, we introduce PersonaLens, a comprehensive benchmark for evalu…

Dringende oproep om ‘onzinnige’ regels rond nieuwbouw te schrappen: ‘We slaan door’ | Wonen | #gelderlander
https://www.gelderlander.nl/wo…