Tootfinder

#PhantastikPrompts 13.10.: No-Gos: Welche Fehler sollten Neu-Autor*innen unbedingt vermeiden?
Ich verstehe die Frage anders als alle anderen, für mich impliziert "No-Go" nicht Fehler, die passieren (wie auf einen DKZV reinfallen), sondern mit Vorsatz Mist bauen, und hinterher unten durch zu sein.
Plagiat und KI-generierte Texte - glauben, dass das eigene Schreiben…

To Ask or Not to Ask: Learning to Require Human Feedback
Andrea Pugnana, Giovanni De Toni, Cesare Barbera, Roberto Pellungrini, Bruno Lepri, Andrea Passerini
https://arxiv.org/abs/2510.08314

To Ask or Not to Ask: Learning to Require Human Feedback
Developing decision-support systems that complement human performance in classification tasks remains an open challenge. A popular approach, Learning to Defer (LtD), allows a Machine Learning (ML) model to pass difficult cases to a human expert. However, LtD treats humans and ML models as mutually exclusive decision-makers, restricting the expert contribution to mere predictions. To address this limitation, we propose Learning to Ask (LtA), a new framework that handles both when and how to inco…

Towards Reliable Retrieval in RAG Systems for Large Legal Datasets
Markus Reuter, Tobias Lingenberg, R\=uta Liepi\c{n}a, Francesca Lagioia, Marco Lippi, Giovanni Sartor, Andrea Passerini, Burcu Sayin
https://arxiv.org/abs/2510.06999

Towards Reliable Retrieval in RAG Systems for Large Legal Datasets
Retrieval-Augmented Generation (RAG) is a promising approach to mitigate hallucinations in Large Language Models (LLMs) for legal applications, but its reliability is critically dependent on the accuracy of the retrieval step. This is particularly challenging in the legal domain, where large databases of structurally similar documents often cause retrieval systems to fail. In this paper, we address this challenge by first identifying and quantifying a critical failure mode we term Document-Leve…

RE: https://todon.nl/@radicalpast/115593474270996660
Wir waren entsetzt, geschockt, sehr wütend auch, dass sowas mitten in Berlin passieren konnte, gleich um die Ecke.
Die Demo war anders als andere Demos.

Wenn jemand so versessen darauf ist, ein Tier zu töten, daß du ihn erschießen mußt, weil er bei dem Versuch dich und andere in Gefahr bringt, dann hol das Tier danach zu dir rein und laß es frei herumlaufen, was soll schon passieren?

Charged Black-Hole Binary Evolution at Second Post-Newtonian Order
Andrea Placidi, Elisa Grilli, Marta Orselli, Matteo Pegorin, Nicola Bartolo, Pierpaolo Mastrolia
https://arxiv.org/abs/2509.20432

Charged Black-Hole Binary Evolution at Second Post-Newtonian Order
We study the dynamics of electrically charged black-hole binaries and their gravitational-wave emission during the inspiral phase. Within the post-Newtonian framework, we derive the conservative and dissipative dynamics up to second order (2PN), combining Effective Field Theory and classical methods. We compute the NNLO conservative Lagrangian, LO dissipative effects in harmonic and Lorenz gauges, and provide the equations of motion, center-of-mass transformations, and the Lagrangian/Hamiltonia…

symptomatiskt https://www.gp.se/kultur/kulturdebatt/utan-kontext-blir-antisemitismens-historia-ett-politiskt-verktyg.68d9fa96-afdc-4f44-b819-8f95055f6b42

Utan kontext blir antisemitismens historia ett politiskt verktyg
Det är möjligt att tala om antisemitism, pogromer och fördrivningar av judar – och samtidigt erkänna den roll sionismen och kolonialmakterna spelade i att forma den verkligheten. Att hävda motsatsen är att göra judisk historia till ett verktyg för politiska syften, skriver Edgar Mannheimer i sin slutreplik till Daniel Leviathan, Isak Nachman och Lizzie Oved Scheja.

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[11/11]:
- Hyperdimensional Probe: Decoding LLM Representations via Vector Symbolic Architectures
Marco Bronzini, Carlo Nicolini, Bruno Lepri, Jacopo Staiano, Andrea Passerini

Hyperdimensional Probe: Decoding LLM Representations via Vector Symbolic Architectures
Marco Bronzini, Carlo Nicolini, Bruno Lepri, Jacopo Staiano, Andrea Passerini
https://arxiv.org/abs/2509.25045

Hyperdimensional Probe: Decoding LLM Representations via Vector Symbolic Architectures
Despite their capabilities, Large Language Models (LLMs) remain opaque with limited understanding of their internal representations. Current interpretability methods, such as direct logit attribution (DLA) and sparse autoencoders (SAEs), provide restricted insight due to limitations such as the model's output vocabulary or unclear feature names. This work introduces Hyperdimensional Probe, a novel paradigm for decoding information from the LLM vector space. It combines ideas from symbolic repre…