Tootfinder

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

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:58:19

AI-Driven Post-Quantum Cryptography for Cyber-Resilient V2X Communication in Transportation Cyber-Physical Systems
Akid Abrar, Sagar Dasgupta, Mizanur Rahman, Ahmad Alsharif
arxiv.org/abs/2510.08496

AI-Driven Post-Quantum Cryptography for Cyber-Resilient V2X Communication in Transportation Cyber-Physical Systems
Transportation Cyber-Physical Systems (TCPS) integrate physical elements, such as transportation infrastructure and vehicles, with cyber elements via advanced communication technologies, allowing them to interact seamlessly. This integration enhances the efficiency, safety, and sustainability of transportation systems. TCPS rely heavily on cryptographic security to protect sensitive information transmitted between vehicles, transportation infrastructure, and other entities within the transporta…

@CondeChocula@social.linux.pizza
2025-11-09 13:28:46

#nostalgia

@randy_@social.linux.pizza
2025-12-08 12:47:14

Swedish Winter really plays with your energy levels.
#sweden #winter #scandinavia

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:56:49

Rethinking Provenance Completeness with a Learning-Based Linux Scheduler
Jinsong Mao, Benjamin E. Ujcich, Shiqing Ma
arxiv.org/abs/2510.08479

Rethinking Provenance Completeness with a Learning-Based Linux Scheduler
Provenance plays a critical role in maintaining traceability of a system's actions for root cause analysis of security threats and impacts. Provenance collection is often incorporated into the reference monitor of systems to ensure that an audit trail exists of all events, that events are completely captured, and that logging of such events cannot be bypassed. However, recent research has questioned whether existing state-of-the-art provenance collection systems fail to ensure the security guar…

@CondeChocula@social.linux.pizza
2025-11-08 15:05:01

Happy #caturday

A cat touching a xbox 360 controller it simulates playing a game.
@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:46:49

ExPrESSO: Zero-Knowledge backed Extensive Privacy Preserving Single Sign-on
Kaustabh Barman, Fabian Piper, Sanjeet Raj Pandey, Axel Kuepper
arxiv.org/abs/2510.08355

ExPrESSO: Zero-Knowledge backed Extensive Privacy Preserving Single Sign-on
User authentication is one of the most important aspects for secure communication between services and end-users over the Internet. Service providers leverage Single-Sign On (SSO) to make it easier for their users to authenticate themselves. However, standardized systems for SSO, such as OIDC, do not guarantee user privacy as identity providers can track user activities. We propose a zero-knowledge-based mechanism that integrates with OIDC to let users authenticate through SSO without revealing…

@dawid@social.craftknight.com
2026-01-08 10:44:56
Po długiej drodze zrobiliśmy sobie przystanek na południowo wschodnim krańcu Sycylii. Rozpogodziło się w ostatnich dniach, ale przyszedł chłodny wiatr i temperatury spadły do jakiś 15 stopni (odczuwalne niby 7).

Przeczekamy tu parę dni, bo jest taka niepisana zasada, że jak znajdzie się dobre miejsce, to trzeba stanąć i korzystać, bo nie wiadomo kiedy kolejne takie się trafi. Niedaleko jest rezerwat, więc jest gdzie chodzić na spacery.

Co rano przyjeżdża ktoś sprzedawać pieczywo, …
@Techmeme@techhub.social
2026-01-08 14:16:28

Illicit crypto addresses received a record $154B in 2025, up 162% YoY, driven by sanctions evasion; Russia's A7A5 token transacted $93.3B in less than one year (Chainalysis)
chainalysis.com/blog/2026-cryp
<…

2026 Crypto Crime Report Introduction - Chainalysis
The professionalization of crypto crime has accelerated, with specialized networks offering laundering-as-a-service to facilitate on- and off-chain crime.

@netzschleuder@social.skewed.de
2025-12-10 00:00:04

us_congress: US Congressional co-sponsorship (1973-2016)
Two temporal networks of bill co-sponsorship tendencies among US Congresspeople, from 1973 (93rd Congress) to 2016 (114th Congress). Edges are signed, indicating the presence of a significant tendency to co-sponsor, or tendency to not co-sponsor, bills. Two networks per year, one for each chamber (House, Senate). Each file contains the signed network for a chamber of congress (H = House, S = Senate) in one session (e.g. 93 = 93rd…

us_congress: US Congressional co-sponsorship (1973-2016). 444 nodes, 66277 edges. https://networks.skewed.de/net/us_congress#H108

us_congress — US Congressional co-sponsorship (1973-2016)
Two temporal networks of bill co-sponsorship tendencies among US Congresspeople, from 1973 (93rd Congress) to 2016 (114th Congress). Edges are signed, indicating the presence of a significant tendency to co-sponsor, or tendency to not co-sponsor, bills. Two networks per year, one for each chamber (House, Senate). Each file contains the signed network for a chamber of congress (H = House, S = Senate) in one session (e.g. 93 = 93rd session of congress).

@phpmacher@sueden.social
2025-11-08 17:57:32

Tag 93: #rausgeschafft
#Samstag

Eine Nachtszene mit silhouettenförmigen Ästen im Vordergrund und beleuchteten Strukturen in der Ferne, teilweise durch Nebel verdeckt.
@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:45:49

A Haskell to FHE Transpiler
Anne M\"uller, Mohd Kashif, Nico D\"ottling
arxiv.org/abs/2510.08343 arxiv.org/pdf/2510.08343

A Haskell to FHE Transpiler
Fully Homomorphic Encryption (FHE) enables the evaluation of programs directly on encrypted data. However, because only basic operations can be performed on ciphertexts, programs must be expressed as boolean or arithmetic circuits. This low-level repre- sentation makes implementing applications for FHE significantly more cumbersome than writing code in a high-level language. To reduce this burden, several transpilers have been developed that translate high-level code into circuit representation…

@mgorny@pol.social
2025-11-10 05:55:47

Wiem, że bardzo was to ciekawi, więc wygrzebałem, że #KolejeWielkopolskie zaczęły honorować sieciówki PR gdzieś pomiędzy listopadem 2021, a styczniem 2022.

Honorowanie biletów - Koleje Wielkopolskie Sp. z o.o.
Honorowanie biletów POLREGIO w Kolejach Wielkopolskich Pociągami Kolei Wielkopolskich można podróżować na podstawie biletów: Jednorazowych według taryfy normalnej z ulgą handlową 25%, 50% z ulgami ustawowymi (33%, 37%, 49%, 51%, 78%, 93%, 95%, 100%) z ulgami 80% i 99% przysługującymi na podstawie odrębnie zawartych umów dla uczestników przejazdów grupowych według taryfy RAZEM – w ofercie […]

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:42:39

New Machine Learning Approaches for Intrusion Detection in ADS-B
Mika\"ela Ngambo\'e, Jean-Simon Marrocco, Jean-Yves Ouattara, Jos\'e M. Fernandez, Gabriela Nicolescu
arxiv.org/abs/2510.08333

New Machine Learning Approaches for Intrusion Detection in ADS-B
With the growing reliance on the vulnerable Automatic Dependent Surveillance-Broadcast (ADS-B) protocol in air traffic management (ATM), ensuring security is critical. This study investigates emerging machine learning models and training strategies to improve AI-based intrusion detection systems (IDS) for ADS-B. Focusing on ground-based ATM systems, we evaluate two deep learning IDS implementations: one using a transformer encoder and the other an extended Long Short-Term Memory (xLSTM) network…

@BBC3MusicBot@mastodonapp.uk
2026-01-09 22:34:09

🇺🇦 #NowPlaying on BBCRadio3's #LateJunction
响马 Bande:
🎵 Toil
#响马Bande

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:41:59

Systematic Assessment of Cache Timing Vulnerabilities on RISC-V Processors
C\'edrick Austa, Jan Tobias M\"uhlberg, Jean-Michel Dricot
arxiv.org/abs/2510.08272

Systematic Assessment of Cache Timing Vulnerabilities on RISC-V Processors
While interest in the open RISC-V instruction set architecture is growing, tools to assess the security of concrete processor implementations are lacking. There are dedicated tools and benchmarks for common microarchitectural side-channel vulnerabilities for popular processor families such as Intel x86-64 or ARM, but not for RISC-V. In this paper we describe our efforts in porting an Intel x86-64 benchmark suite for cache-based timing vulnerabilities to RISC-V. We then use this benchmark to eva…

@randy_@social.linux.pizza
2025-11-06 14:18:43

What a nice day! Sun is shining and we have a salary increase! :amongustwerk: :frog_glasses: :data_laughing:
#sunny #happythursday #greatday

@NFL@darktundra.xyz
2025-12-09 20:09:34

Ex-CB Jammer: Played NFL games while drunk espn.com/nfl/story/_/id/472583

Ex-Chargers CB Quentin Jammer says he played while drunk - ESPN
Former San Diego Chargers cornerback Quentin Jammer posted to social media that he played several games while intoxicated during his NFL career.

@tagesschau_eil@mastodon.social
2025-12-06 18:45:27

📢 De Masi und Mohamed Ali neues BSW-Führungsduo
Der Europaabgeordnete Fabio De Masi ist zum neuen Vorsitzenden des BSW gewählt worden. Der 45-Jährige erhielt beim Bundesparteitag in Magdeburg nach Angaben des Tagungsleiters 93,3 Prozent der abgegebenen Stimmen. Die bisherige Co-Vorsitzende Amira Mohamed Ali wurde demnach mit 82,6 Prozent der Stimmen im Amt bestätigt.
➡️

De Masi und Mohamed Ali neues BSW-Führungsduo
Der Europaabgeordnete Fabio De Masi ist zum neuen Vorsitzenden des BSW gewählt worden. Der 45-Jährige erhielt beim Bundesparteitag in Magdeburg nach Angaben des Tagungsleiters 93,3 Prozent der abgegebenen Stimmen. Die bisherige Co-Vorsitzende Amira Mohamed Ali wurde demnach mit 82,6 Prozent der Stimmen im Amt bestätigt.

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:40:19

TracE2E: Easily Deployable Middleware for Decentralized Data Traceability
Daniel Pressens\'e, Elisavet Kozyri
arxiv.org/abs/2510.08225 …

TracE2E: Easily Deployable Middleware for Decentralized Data Traceability
This paper presents TracE2E, a middleware written in Rust, that can provide both data explainability and compliance across multiple nodes. By mediating inputs and outputs of processes, TracE2E records provenance information and enforces data-protection policies (e.g., confidentiality, integrity) that depend on the recorded provenance. Unlike existing approaches that necessitate substantial application modifications, TracE2E is designed for easy integration into existing and future applications …

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:34:29

LLM-Assisted Web Measurements
Simone Bozzolan, Stefano Calzavara, Lorenzo Cazzaro
arxiv.org/abs/2510.08101 arxiv.org/pdf/2510.08101

LLM-Assisted Web Measurements
Web measurements are a well-established methodology for assessing the security and privacy landscape of the Internet. However, existing top lists of popular websites commonly used as measurement targets are unlabeled and lack semantic information about the nature of the sites they include. This limitation makes targeted measurements challenging, as researchers often need to rely on ad-hoc techniques to bias their datasets toward specific categories of interest. In this paper, we investigate the…

@usul@piaille.fr
2026-01-07 18:55:24

Flying Drones with the Dnipro Battalion: Russian Roulette (Dispatch 93) - YouTube
#drone

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:31:19

A Novel Ensemble Learning Approach for Enhanced IoT Attack Detection: Redefining Security Paradigms in Connected Systems
Hikmat A. M. Abdeljaber, Md. Alamgir Hossain, Sultan Ahmad, Ahmed Alsanad, Md Alimul Haque, Sudan Jha, Jabeen Nazeer
arxiv.org/abs/2510.08084

A Novel Ensemble Learning Approach for Enhanced IoT Attack Detection: Redefining Security Paradigms in Connected Systems
The rapid expansion of Internet of Things (IoT) devices has transformed industries and daily life by enabling widespread connectivity and data exchange. However, this increased interconnection has introduced serious security vulnerabilities, making IoT systems more exposed to sophisticated cyber attacks. This study presents a novel ensemble learning architecture designed to improve IoT attack detection. The proposed approach applies advanced machine learning techniques, specifically the Extra T…

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:24:19

Composition Law of Conjugate Observables in Random Permutation Sorting Systems
Yurang R. Kuang
arxiv.org/abs/2510.08013 arxiv.org/pdf/2510.…

Composition Law of Conjugate Observables in Random Permutation Sorting Systems
We present the discovery of a fundamental composition law governing conjugate observables in the Random Permutation Sorting System (RPSS). The law links the discrete permutation count Np and the continuous elapsed time T through a functional relation connecting the characteristic function of timing distributions to the probability generating function of permutation counts. This framework enables entropy purification, transforming microarchitectural timing fluctuations into uniform randomness vi…

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:19:19

From Defender to Devil? Unintended Risk Interactions Induced by LLM Defenses
Xiangtao Meng, Tianshuo Cong, Li Wang, Wenyu Chen, Zheng Li, Shanqing Guo, Xiaoyun Wang
arxiv.org/abs/2510.07968

From Defender to Devil? Unintended Risk Interactions Induced by LLM Defenses
Large Language Models (LLMs) have shown remarkable performance across various applications, but their deployment in sensitive domains raises significant concerns. To mitigate these risks, numerous defense strategies have been proposed. However, most existing studies assess these defenses in isolation, overlooking their broader impacts across other risk dimensions. In this work, we take the first step in investigating unintended interactions caused by defenses in LLMs, focusing on the complex in…

@cdarwin@c.im
2025-11-06 20:00:14

IRS Direct File is open source
Here's it's repository c.im/@cdarwin/1154947762082779

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:09:19

Decentralised Blockchain Management Through Digital Twins
Georgios Diamantopoulos, Nikos Tziritas, Rami Bahsoon, Georgios Theodoropoulos
arxiv.org/abs/2510.07901

Decentralised Blockchain Management Through Digital Twins
The necessity of blockchain systems to remain decentralised limits current solutions to blockchain governance and dynamic management, forcing a trade-off between control and decentralisation. In light of the above, this work proposes a dynamic and decentralised blockchain management mechanism based on digital twins. To ensure decentralisation, the proposed mechanism utilises multiple digital twins that the system's stakeholders control. To facilitate decentralised decision-making, the twins are…

@NFL@darktundra.xyz
2025-11-09 12:24:43

'There's an energy and a competitive spirit to Stef': Why Diggs has boosted Patriots espn.com/nfl/story/_/id/468860

Patriots' Stefon Diggs hits milestone, praised for leadership - ESPN
Receiver Stefon Diggs earns high marks from coaches and teammates for his leadership and desire to win.

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:08:59

Effective and Stealthy One-Shot Jailbreaks on Deployed Mobile Vision-Language Agents
Renhua Ding, Xiao Yang, Zhengwei Fang, Jun Luo, Kun He, Jun Zhu
arxiv.org/abs/2510.07809

Effective and Stealthy One-Shot Jailbreaks on Deployed Mobile Vision-Language Agents
Large vision-language models (LVLMs) enable autonomous mobile agents to operate smartphone user interfaces, yet vulnerabilities to UI-level attacks remain critically understudied. Existing research often depends on conspicuous UI overlays, elevated permissions, or impractical threat models, limiting stealth and real-world applicability. In this paper, we present a practical and stealthy one-shot jailbreak attack that leverages in-app prompt injections: malicious applications embed short prompts…

@wischekurier@mastodon.social
2025-10-10 07:36:02

aktuelles #Wetter #Behrendorf #Wische 10-10-2025 09:30
Temperatur: (min: 11,7C) 13,0C (max: 13,0C)
Feuchte: 93%
Luftdruck: 1026,6 hPa (stark steigend)
Windrichtung: 296 Grad
Wi…

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 09:08:09

ANCORA: Accurate Intrusion Recovery for Web Applications
Yihao Peng, Biao Ma, Hai Wan, Xibin Zhao
arxiv.org/abs/2510.07806 arxiv.org/pdf/25…

ANCORA: Accurate Intrusion Recovery for Web Applications
Modern web application recovery presents a critical dilemma. Coarse-grained snapshot rollbacks cause unacceptable data loss for legitimate users. Surgically removing an attack's impact is hindered by a fundamental challenge in high-concurrency environments: it is difficult to attribute resulting file and database modifications to a specific attack-related request. We present ANCORA, a system for precise intrusion recovery in web applications without invasive instrumentation. ANCORA first isolat…

@samvarma@fosstodon.org
2025-11-06 14:43:56

#GoodNews @… wandering.shop/@adapalmer/1154

Ada Palmer (@adapalmer@wandering.shop)
The World Bank’s latest data shows a quiet global triumph: 93% of people aged 15 to 24 can now read and write, up from 87% in 2000. In many regions, including East Asia, Latin America, and Central Europe, youth literacy has reached or neared universal levels, marking one of the most successful, least reported stories in development. https://data.worldbank.org/indicator/SE.ADT.1524.LT.ZS?end=2024&start=2000&utm_source=substack&utm_medium=email #ShareGoodNewsToo

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:58:58

Rethinking Reasoning: A Survey on Reasoning-based Backdoors in LLMs
Man Hu, Xinyi Wu, Zuofeng Suo, Jinbo Feng, Linghui Meng, Yanhao Jia, Anh Tuan Luu, Shuai Zhao
arxiv.org/abs/2510.07697

Rethinking Reasoning: A Survey on Reasoning-based Backdoors in LLMs
With the rise of advanced reasoning capabilities, large language models (LLMs) are receiving increasing attention. However, although reasoning improves LLMs' performance on downstream tasks, it also introduces new security risks, as adversaries can exploit these capabilities to conduct backdoor attacks. Existing surveys on backdoor attacks and reasoning security offer comprehensive overviews but lack in-depth analysis of backdoor attacks and defenses targeting LLMs' reasoning abilities. In this…

@noellabo@fedibird.com
2025-12-04 11:54:58

まあ、松井秀喜よね
#このタグを見た人は自分と同じ誕生日の人を1人言う

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:55:39

A Minrank-based Encryption Scheme \`a la Alekhnovich-Regev
Thomas Debris-Alazard, Philippe Gaborit, Romaric Neveu, Olivier Ruatta
arxiv.org/abs/2510.07584

A Minrank-based Encryption Scheme à la Alekhnovich-Regev
Introduced in 2003 and 2005, Alekhnovich and Regev' schemes were the first public-key encryptions whose security is only based on the average hardness of decoding random linear codes and LWE, without other security assumptions. Such security guarantees made them very popular, being at the origin of the now standardized HQC or Kyber. We present an adaptation of Alekhnovich and Regev' encryption scheme whose security is only based on the hardness of a slight variation of MinRank, the so-called …

@raiders@darktundra.xyz
2025-10-27 12:04:45

‘First Lady of Raider Nation’ dies at age 93 reviewjournal.com/sports/raide

‘First Lady of Raider Nation’ dies at age 93
Carol Davis, the mother of Raiders owner Mark Davis and the husband of late Raiders owner Al Davis, died Friday at age 93.

@BBC6MusicBot@mastodonapp.uk
2026-01-08 08:45:04

🇺🇦 #NowPlaying on #BBC6Music's #NickGrimshaw
Souls of Mischief:
🎵 93 Til Infinity
#SoulsofMischief
lurre.bandcamp.com/track/souls
open.spotify.com/track/0PV1TFU
Please 🔁 BOOST to share what you like
- your followers don't see if you ⭐ favourite a post

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:49:58

EMPalm: Exfiltrating Palm Biometric Data via Electromagnetic Side-Channels
Haowen Xu, Tianya Zhao, Xuyu Wang, Lei Ma, Jun Dai, Alexander Wyglinski, Xiaoyan Sun
arxiv.org/abs/2510.07533

EMPalm: Exfiltrating Palm Biometric Data via Electromagnetic Side-Channels
Palm recognition has emerged as a dominant biometric authentication technology in critical infrastructure. These systems operate in either single-modal form, using palmprint or palmvein individually, or dual-modal form, fusing the two modalities. Despite this diversity, they share similar hardware architectures that inadvertently emit electromagnetic (EM) signals during operation. Our research reveals that these EM emissions leak palm biometric information, motivating us to develop EMPalm--an a…

@nohillside@smnn.ch
2026-01-06 06:43:05

“#pastpuzzle 209
🟩🟥🟩🟥 (-93)
🟩🟩🟥🟥 (-17)
🟩🟩🟥🟥 (-39)
🟩🟩🟩🟥 (-1)
x/4 🟥” pastpuzzle.de/

past puzzle
Errate das gesuchte Jahr mit Hilfe von 4 historischen Ereignissen. Ein von Wordle und Geschichte inspiriertes Spiel.

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:46:19

MIRANDA: short signatures from a leakage-free full-domain-hash scheme
Alain Couvreur, Thomas Debris-Alazard, Philippe Gaborit, Adrien Vin\c{c}otte
arxiv.org/abs/2510.07479

MIRANDA: short signatures from a leakage-free full-domain-hash scheme
We present $\mathsf{Miranda}$, the first family of full-domain-hash signatures based on matrix codes. This signature scheme fulfils the paradigm of Gentry, Peikert and Vaikuntanathan ($\mathsf{GPV}$), which gives strong security guarantees. Our trapdoor is very simple and generic: if we propose it with matrix codes, it can actually be instantiated in many other ways since it only involves a subcode of a decodable code (or lattice) in a unique decoding regime of parameters. Though $\mathsf{Miran…

@mlippert@vmst.io
2025-11-06 17:06:27

#Wordle 1,601 5/6*
⬜🟨⬜⬜⬜ <1% of 229,282 (277)
⬜⬜🟩⬜⬜ 0 of 31 (34)
🟨⬜🟩⬜🟩 0 of 0 (2)
⬜⬜🟩🟩🟩 0 of 0 (1)
🟩🟩🟩🟩🟩
WordleBot
Skill 93/99
Luck 35/99
I liked today's word, and despite not being the right word, I liked my 4th guess too and I'm glad I made it!

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:36:48

A Secure Authentication-Driven Protected Data Collection Protocol in Internet of Things
Maryam Ataei Nezhad, Hamid Barati, Ali Barati
arxiv.org/abs/2510.07462

A Secure Authentication-Driven Protected Data Collection Protocol in Internet of Things
Internet of Things means connecting different devices through the Internet. The Internet of things enables humans to remotely manage and control the objects they use with the Internet infrastructure. After the advent of the Internet of Things in homes, organizations, and private companies, privacy and information security are the biggest concern. This issue has challenged the spread of the Internet of things as news of the users theft of information by hackers intensified. The proposed method i…

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:34:39

Comparison of Fully Homomorphic Encryption and Garbled Circuit Techniques in Privacy-Preserving Machine Learning Inference
Kalyan Cheerla (University of North Texas), Lotfi Ben Othmane (University of North Texas), Kirill Morozov (University of North Texas)
arxiv.org/abs/2510.07457

Comparison of Fully Homomorphic Encryption and Garbled Circuit Techniques in Privacy-Preserving Machine Learning Inference
Machine Learning (ML) is making its way into fields such as healthcare, finance, and Natural Language Processing (NLP), and concerns over data privacy and model confidentiality continue to grow. Privacy-preserving Machine Learning (PPML) addresses this challenge by enabling inference on private data without revealing sensitive inputs or proprietary models. Leveraging Secure Computation techniques from Cryptography, two widely studied approaches in this domain are Fully Homomorphic Encryption (F…

@cheeaun@mastodon.social
2025-10-27 04:06:38

RE: mastodon.social/@cheeaun/11541
After looking at this, got curious to know the limits in most servers.
So I did a little data analysis. Servers list from @…

Chart titled "Image Matrix Limits" showing a table lists matrix MP values (2, 17, 33, 38–195) with counts and percentages and bar graph: 33 MP dominates (2145, 93.50%), 17 MP (139, 6.06%), and small entries for 2 MP and 38–195 MP.
Chart titled "Image Size Limits" showing counts and percentages of image sizes (MB) with a horizontal bar graph. The 16 MB row dominates (2,081 items, 90.71%) while other size buckets (4–5, 8, 10, 15, 19, 20, 24–32, 38–48, 50–99, 100–1354 MB) show much smaller counts and percentages.
Chart titled "Video Matrix Limits" showing matrix sizes 2MP (138, 6.02%), 8MP (2149, 93.68%) and 9–36MP (7, 0.31%) with horizontal bar graph.
Chart titled "Video Size Limits" showing size bins (10–20, 40, 50–80, 86–98, 99, 100, 128–160, 200, 250–800, 990–2048 MB) with counts and percentages; the 99 MB row dominates with count 2086 (90.93%).
@fanf@mendeddrum.org
2025-11-28 09:42:01

from my link log —
Dumbindent: because 93% of the time was spent in clang-format.
nigeltao.github.io/blog/2020/d
saved 2020-06-15

@andres4ny@social.ridetrans.it
2025-11-05 02:30:18

lmao, if mamdani's lead over cuomo ends up being the same number of votes that sliwa gets...

Totals by Candidate Mamdani: 49.6% Sliwa: 7.93% Cuomo: 41.61%
@arXiv_csCR_bot@mastoxiv.page
2025-10-10 08:26:39

PATCH: Mitigating PII Leakage in Language Models with Privacy-Aware Targeted Circuit PatcHing
Anthony Hughes, Vasisht Duddu, N. Asokan, Nikolaos Aletras, Ning Ma
arxiv.org/abs/2510.07452

PATCH: Mitigating PII Leakage in Language Models with Privacy-Aware Targeted Circuit PatcHing
Language models (LMs) may memorize personally identifiable information (PII) from training data, enabling adversaries to extract it during inference. Existing defense mechanisms such as differential privacy (DP) reduce this leakage, but incur large drops in utility. Based on a comprehensive study using circuit discovery to identify the computational circuits responsible PII leakage in LMs, we hypothesize that specific PII leakage circuits in LMs should be responsible for this behavior. Therefor…

@arXiv_csCR_bot@mastoxiv.page
2025-10-10 07:42:19

[2025-10-10 Fri (UTC), 21 new articles found for cs.CR Cryptography and Security]
toXiv_bot_toot

@zack@mamot.fr
2025-12-18 00:18:22

Noël: une fabuleuse occasion pour enfin faire connaissance avec la fast-déco, prendre conscience de ces impacts,
ou pester qu'elle ait été exclue du récent débat public sur la fast-fashion.
Le cynisme débonnaire de ces grandes enseignes (françaises comme étrangères) me révulse: extractivisme, pollution plastique, dégâts sociaux...
La fast-déco n'a rien Š envier Š sa copine toxique la fast-fashion.
Ouais, parlons-en de la fast-déco.
Parlons-en beaucoup.

Surconsommation : à l’approche de Noël, la fast-déco fout les boules
Nombreux sont les Français à se ruer dans les magasins à la recherche de produits à bas coût. S’ils sont pourtant fabriqués pour la plupart en Asie, des centaines de millions d’objets d’intérieur sont vendus chaque année.

@relcfp@mastodon.social
2025-11-05 05:26:18

Call for Chapters: Bodies That Breathe – The Politics of Air, Health, and Survival (Volume 1 of the “Earth and Us” Book Series) call-for-papers.sas.upenn.edu/

@chrislowles@mastodon.social
2025-11-07 07:01:16

Remember: If you have enough money you can get away with anything npr.org/2025/11/07/nx-s1-56015

@wischekurier@mastodon.social
2026-01-09 04:16:11

aktuelles #Wetter #Behrendorf #Wische 09-01-2026 05:10
Temperatur: (min: -5,0C) -4,8C (max: -4,4C)
Feuchte: 93%
Luftdruck: 1001,6 hPa (schnell fallend)
Windrichtung: 221 Grad
W…

@randy_@social.linux.pizza
2025-10-30 15:01:12

@tekklinik@mastodon.social @…
Ill give Comaps a try, but the metro aspect is not better.

@BBC3MusicBot@mastodonapp.uk
2025-12-06 11:44:08

🇺🇦 #NowPlaying on BBCRadio3's #SaturdayMorning
Ludwig van Beethoven, Nikolaus Harnoncourt, Chamber Orchestra of Europe, Charlotte Margiono, Rudolf Schasching, Birgit Remmert, Robert Holl, Arnold Schoenberg Chor & Erwin Ortner:
🎵 Symphony No 8 in F major, Op 93 (2nd mvt)
open.spotify.com/track/18LVcYN

@raiders@darktundra.xyz
2025-10-26 22:39:16

‘First Lady of Raider Nation’ dies at age 93 reviewjournal.com/sports/raide

‘First Lady of Raider Nation’ dies at age 93
Carol Davis, the mother of Raiders owner Mark Davis and the husband of late Raiders owner Al Davis, died Friday at age 93.

@NFL@darktundra.xyz
2025-12-30 07:00:43

Bijan Robinson soars to 93-yard TD in stellar showing vs. Rams: I was like, 'I've got to take off' nfl.com/news/bijan-robinson-so

Bijan Robinson soars to 93-yard TD in stellar showing vs. Rams: I was like, 'I've got to take off'
This has been a trying season for the Atlanta Falcons, who have fallen well short of expectations. The same cannot be said for Bijan Robinson, a one-man highlight reel in 2025, who befuddled the Los Angeles Rams and dazzled viewers on Monday night.

@mlippert@vmst.io
2025-12-06 17:58:24

#Wordle 1,631 3/6*
⬜⬜⬜🟨🟩 <1% of 232,908 (50)
⬜⬜🟨⬜🟩 11% of 44 (8)
🟩🟩🟩🟩🟩 33% of 6
WordleBot
Skill 83/99
Luck 67/99
Really thought there were lots of possible words left, and my 3rd choice wasn't likely. Bot agrees.
3rd guess lowered my skill a bunch with a 66 and raised my luck a lot with a 93.

@wischekurier@mastodon.social
2026-01-09 03:16:13

aktuelles #Wetter #Behrendorf #Wische 09-01-2026 04:10
Temperatur: (min: -5,0C) -4,8C (max: -4,4C)
Feuchte: 93%
Luftdruck: 1002,6 hPa (schnell fallend)
Windrichtung: 124 Grad
W…

@Techmeme@techhub.social
2025-12-02 18:05:47

Filings: after EA's buyout, Saudi Arabia's PIF would own 93.4% of EA, implying it needs to put up ~$29B, while Silver Lake and Affinity would own 5.5% and 1.1% (Wall Street Journal)
wsj.com/business/deals/saudi-f

@randy_@social.linux.pizza
2025-10-30 06:37:26

I’ve been testing @… for the past week or so. I’ve been leaving the house with mobile data turned off, relying only on Organic Maps for directions. Saving locations, using offline maps—everything works as smooth as butter, and I haven’t noticed any issues with battery life. I’d leave with 100% and come back with 94%, even after turning data on a few times t…

@CondeChocula@social.linux.pizza
2025-11-06 20:12:26

One of the best console in my opinion. :n64_controller:
#retro #retrogaming #nintendo #nintendo64

Nintendo Magazine UK #54, March 1997 - 'Super Mario 64' cover.
@zack@mamot.fr
2025-12-16 00:53:43

"Le cas du député Emmanuel Pellerin révèle les contradictions entre le discours politique intransigeant Š l’égard de la consommation de drogues et la réalité de la réponse judiciaire quand un élu est mis cause pour de tels faits" résume le tribunal dans son jugement, en indiquant que ces révélations portent pleinement sur un « sujet d’intérêt général ».
C'est compliqué, apparemment, pour un avocat-député de comprendre cette notion d'intérêt général.

Cocaïne en politique : l’ex-député Pellerin condamné pour procédure abusive contre Mediapart
Le tribunal de Paris a relaxé Mediapart du procès en diffamation que lui avait intenté l’ancien élu Renaissance Emmanuel Pellerin. Il a même condamné l’ex-député, avocat de profession, au paiement de dommages et intérêts.

@NFL@darktundra.xyz
2026-01-07 20:59:48

Jerry Jones: Goal is to retire with most SB wins espn.com/nfl/story/_/id/475416

Cowboys' Jerry Jones: Goal is to retire with most Super Bowl wins - ESPN
Cowboys owner Jerry Jones said Wednesday that his goal in life is to the retire as the owner with the most Super Bowl wins and that he has "work to do" to catch the Patriots' Robert Kraft.

@netzschleuder@social.skewed.de
2025-12-03 04:00:05

us_congress: US Congressional co-sponsorship (1973-2016)
Two temporal networks of bill co-sponsorship tendencies among US Congresspeople, from 1973 (93rd Congress) to 2016 (114th Congress). Edges are signed, indicating the presence of a significant tendency to co-sponsor, or tendency to not co-sponsor, bills. Two networks per year, one for each chamber (House, Senate). Each file contains the signed network for a chamber of congress (H = House, S = Senate) in one session (e.g. 93 = 93rd…

us_congress: US Congressional co-sponsorship (1973-2016). 447 nodes, 63851 edges. https://networks.skewed.de/net/us_congress#H107

us_congress — US Congressional co-sponsorship (1973-2016)
Two temporal networks of bill co-sponsorship tendencies among US Congresspeople, from 1973 (93rd Congress) to 2016 (114th Congress). Edges are signed, indicating the presence of a significant tendency to co-sponsor, or tendency to not co-sponsor, bills. Two networks per year, one for each chamber (House, Senate). Each file contains the signed network for a chamber of congress (H = House, S = Senate) in one session (e.g. 93 = 93rd session of congress).

@wischekurier@mastodon.social
2026-01-08 11:11:12

aktuelles #Wetter #Behrendorf #Wische 08-01-2026 12:00
Temperatur: (min: -10,3C) -5,0C (max: -5,0C)
Feuchte: 93%
Luftdruck: 1008,6 hPa (gleichbleibend)
Windrichtung: 197 Grad
W…

@raiders@darktundra.xyz
2025-10-26 23:03:58

The First Lady of Raider Nation, Carol Davis, has passed away at 93 raiders.com/video/the-first-la

The First Lady of Raider Nation, Carol Davis, has passed away at 93
Earlier today at 1:05pm, Mark Davis lit the Al Davis Memorial Torch at Allegiant Stadium in honor of and tribute to his mother, Carol Davis, the First Lady of Raider Nation, who passed away on Friday afternoon at the age of 93.

@randy_@social.linux.pizza
2025-10-29 06:35:30

Bank account: "Did you just book hotels in Japan?"
Me: "oh i did"
Seems like I'm going to japan, next year!
#japan #travel #asia

@wischekurier@mastodon.social
2026-01-08 10:25:53

aktuelles #Wetter #Behrendorf #Wische 08-01-2026 11:15
Temperatur: (min: -10,3C) -5,2C (max: -5,0C)
Feuchte: 93%
Luftdruck: 1008,9 hPa (gleichbleibend)
Windrichtung: 195 Grad
W…

@raiders@darktundra.xyz
2025-10-26 23:40:43

Carol Davis, 'First Lady of Raider Nation,' passes away at 93 raiderswire.usatoday.com/story

Carol Davis, 'First Lady of Raider Nation,' passes away at 93
Carol Davis, co-owner of the Las Vegas Raiders and wife of the late Al Davis, passed away Friday at age 93.

@mlippert@vmst.io
2025-12-05 16:43:28

#Wordle 1,630 4/6*
⬜🟨⬜⬜⬜ >1% of 228,180 (164)
⬜⬜🟩⬜⬜ 4% of 110 (14)
🟩🟨🟩⬜⬜ 0 of 4 (2)
🟩🟩🟩🟩🟩
WordleBot
Skill 93/99
Luck 45/99
I was pretty happy with my guesses today. Some days it's hard to find a word that tests what you want to test, but today I found words that did what I wanted :-)

@randy_@social.linux.pizza
2025-10-29 03:24:59

Fuck! I'm out of coffee. The end is near!
#coffee #drama

@zack@mamot.fr
2025-12-16 00:46:33

A ne pas confondre avec un rhume de cerveau qui nécessite toi-même-tu-sais-hein.
liberation.fr/politique/franco

François Bayrou hospitalisé pour «une grippe très sévère»
Le cabinet du maire de Pau a indiqué ce lundi que l’ex-premier ministre avait été transféré aux urgences du centre hospitalier de la capitale béarnaise.

@raiders@darktundra.xyz
2025-10-26 23:03:59

Carol Davis, widow of Al Davis and 'First Lady of Raider Nation,' dies at 93 foxsports.com/articles/nfl/car

Carol Davis, widow of Al Davis and 'First Lady of Raider Nation,' dies at 93
Carol Davis, the widow of longtime Raiders owner Al Davis, has died at 93

@CondeChocula@social.linux.pizza
2025-11-16 14:18:03

Today, channel "Cuatro" from spain is about to stream the first live game of NFL in Spain (Santiago Bernabeu).
I don't have any idea of the rules of the Game 😅
Never I have seen a match from NFL
#nfl #dolphins

@randy_@social.linux.pizza
2025-12-26 12:37:36

The season of acting like we are all catholic and care about Jesus is over. We can stop saying Merry Christmas for the 100th time to the same person.
#christmas2025

@zack@mamot.fr
2025-12-16 00:34:44

* L'État doit-il protéger les consommateur.ice.s, ou seulement assurer ses missions régaliennes ?
* Un service public se mesure-t-il Š son utilité sociale ou Š son équilibre budgétaire ?
* Une information indépendante sous financement privé est-elle viable?
Un choix politique, deux modèles de société... La Cour des Comptes, le gouvernement Lecornu et le Parlement ont choisi.
"Un bon service public est un service public mort."

Le Sénat acte la liquidation de « 60 Millions de Consommateurs », une mauvaise nouvelle pour les citoyens
L’Institut national de la Consommation et son journal 60 Millions de Consommateurs devraient être dissous dans les semaines à venir, a ordonné le Sénat.

@randy_@social.linux.pizza
2025-11-25 08:10:54

It's Kimchi making day.
#Kimchi #food #KoreanFood

@CondeChocula@social.linux.pizza
2025-11-15 14:59:57

Thinking to do a run Baldur's Gate Enhanced Edition on Steam...
Years ago I played but never ended it...
#baldursgate #retrogame #retrogaming

@raiders@darktundra.xyz
2025-10-26 21:56:18

The First Lady of Raider Nation, Carol Davis, has passed away at 93 raiders.com/news/carol-davis-t

@NFL@darktundra.xyz
2025-11-05 22:24:37

MVS, Rodgers all smiles after again joining forces espn.com/nfl/story/_/id/468620

Steelers' Marquez Valdes-Scantling, Aaron Rodgers happy with reunion - ESPN
WR Marquez Valdes-Scantling, who spent four seasons with Aaron Rodgers in Green Bay, is excited about reconnecting with one of his "closest friends" in Pittsburgh.

@randy_@social.linux.pizza
2025-11-20 15:23:07

Winter.
Temperatures start dropping below freezing. Christmas lights pop up like mushrooms.
The first snow is on the way, and the smell of burning wood becomes the new normal.
Winter has arrived.
#winter #blog

@randy_@social.linux.pizza
2025-10-17 12:56:58

Done.
From: @…
mastodon.social/@Tutanota/1153

Tuta (@Tutanota@mastodon.social)
Attached: 1 image 🚨⏳ Act now! You have until November 3rd, 2025 to prevent Microsoft from using your LinkedIn data to train AI. You're opted in automatically, unless you take action and turn it off. Here's how: Go to your account → Settings & Privacy → Data privacy → Data for Generative AI Improvement & toggle the switch off to opt out. Find out more: https://tuta.com/blog/linkedin-ai-user-data #LinkedInAI #MicrosoftAI #BigTech

@randy_@social.linux.pizza
2025-10-16 16:11:26

Always something.
Sourdough bread with brie, a big coffee with good company. That’s how I started my morning: breakfast in a French bakery in Poland.
Next to us, a lady who takes French breakfast to a Slavic level in a fancy way: coupe champagne, assortment bread, lightly salted butter, and cheese—but the cigarette was missing. Actually, what she had was actually what I wanted. After saying goodbye to my friend, I headed to find the train station and a supermarket—with success! On…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:26:38

PACEbench: A Framework for Evaluating Practical AI Cyber-Exploitation Capabilities
Zicheng Liu, Lige Huang, Jie Zhang, Dongrui Liu, Yuan Tian, Jing Shao
arxiv.org/abs/2510.11688

PACEbench: A Framework for Evaluating Practical AI Cyber-Exploitation Capabilities
The increasing autonomy of Large Language Models (LLMs) necessitates a rigorous evaluation of their potential to aid in cyber offense. Existing benchmarks often lack real-world complexity and are thus unable to accurately assess LLMs' cybersecurity capabilities. To address this gap, we introduce PACEbench, a practical AI cyber-exploitation benchmark built on the principles of realistic vulnerability difficulty, environmental complexity, and cyber defenses. Specifically, PACEbench comprises four…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:25:38

Bag of Tricks for Subverting Reasoning-based Safety Guardrails
Shuo Chen, Zhen Han, Haokun Chen, Bailan He, Shengyun Si, Jingpei Wu, Philip Torr, Volker Tresp, Jindong Gu
arxiv.org/abs/2510.11570

Bag of Tricks for Subverting Reasoning-based Safety Guardrails
Recent reasoning-based safety guardrails for Large Reasoning Models (LRMs), such as deliberative alignment, have shown strong defense against jailbreak attacks. By leveraging LRMs' reasoning ability, these guardrails help the models to assess the safety of user inputs before generating final responses. The powerful reasoning ability can analyze the intention of the input query and will refuse to assist once it detects the harmful intent hidden by the jailbreak methods. Such guardrails have show…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:24:08

Uncertainty-Aware, Risk-Adaptive Access Control for Agentic Systems using an LLM-Judged TBAC Model
Charles Fleming, Ashish Kundu, Ramana Kompella
arxiv.org/abs/2510.11414

Uncertainty-Aware, Risk-Adaptive Access Control for Agentic Systems using an LLM-Judged TBAC Model
The proliferation of autonomous AI agents within enterprise environments introduces a critical security challenge: managing access control for emergent, novel tasks for which no predefined policies exist. This paper introduces an advanced security framework that extends the Task-Based Access Control (TBAC) model by using a Large Language Model (LLM) as an autonomous, risk-aware judge. This model makes access control decisions not only based on an agent's intent but also by explicitly considerin…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:23:48

Living Off the LLM: How LLMs Will Change Adversary Tactics
Sean Oesch, Jack Hutchins, Luke Koch, Kevin Kurian
arxiv.org/abs/2510.11398 arxi…

Living Off the LLM: How LLMs Will Change Adversary Tactics
In living off the land attacks, malicious actors use legitimate tools and processes already present on a system to avoid detection. In this paper, we explore how the on-device LLMs of the future will become a security concern as threat actors integrate LLMs into their living off the land attack pipeline and ways the security community may mitigate this threat.

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:21:58

TBRD: TESLA Authenticated UAS Broadcast Remote ID
Jason Veara, Manav Jain, Kyle Moy, Aanjhan Ranganathan
arxiv.org/abs/2510.11343 arxiv.org…

TBRD: TESLA Authenticated UAS Broadcast Remote ID
Mysterious sightings of Unmanned Aircraft Systems (UAS) over U.S. military facilities, suburban neighborhoods, and commercial airports have intensified scrutiny of drone activity. To increase accountability, the Federal Aviation Administration (FAA) introduced a Remote ID mandate, requiring unmanned aircraft to broadcast their location, operator's location, and identity in real-time. However, current standards leave authentication mechanisms underspecified, enabling spoofing, relay, and replay …

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:21:28

TDADL-IE: A Deep Learning-Driven Cryptographic Architecture for Medical Image Security
Junhua Zhou, Quanjun Li, Weixuan Li, Guang Yu, Yihua Shao, Yihang Dong, Mengqian Wang, Zimeng Li, Changwei Gong, Xuhang Chen
arxiv.org/abs/2510.11301

TDADL-IE: A Deep Learning-Driven Cryptographic Architecture for Medical Image Security
The rise of digital medical imaging, like MRI and CT, demands strong encryption to protect patient data in telemedicine and cloud storage. Chaotic systems are popular for image encryption due to their sensitivity and unique characteristics, but existing methods often lack sufficient security. This paper presents the Three-dimensional Diffusion Algorithm and Deep Learning Image Encryption system (TDADL-IE), built on three key elements. First, we propose an enhanced chaotic generator using an LST…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:19:48

How to Get Actual Privacy and Utility from Privacy Models: the k-Anonymity and Differential Privacy Families
Josep Domingo-Ferrer, David S\'anchez
arxiv.org/abs/2510.11299

How to Get Actual Privacy and Utility from Privacy Models: the k-Anonymity and Differential Privacy Families
Privacy models were introduced in privacy-preserving data publishing and statistical disclosure control with the promise to end the need for costly empirical assessment of disclosure risk. We examine how well this promise is kept by the main privacy models. We find they may fail to provide adequate protection guarantees because of problems in their definition or incur unacceptable trade-offs between privacy protection and utility preservation. Specifically, k-anonymity may not entirely exclude …

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:17:58

Large Language Models Are Effective Code Watermarkers
Rui Xu, Jiawei Chen, Zhaoxia Yin, Cong Kong, Xinpeng Zhang
arxiv.org/abs/2510.11251 a…

Large Language Models Are Effective Code Watermarkers
The widespread use of large language models (LLMs) and open-source code has raised ethical and security concerns regarding the distribution and attribution of source code, including unauthorized redistribution, license violations, and misuse of code for malicious purposes. Watermarking has emerged as a promising solution for source attribution, but existing techniques rely heavily on hand-crafted transformation rules, abstract syntax tree (AST) manipulation, or task-specific training, limiting …

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:16:28

Collaborative Shadows: Distributed Backdoor Attacks in LLM-Based Multi-Agent Systems
Pengyu Zhu, Lijun Li, Yaxing Lyu, Li Sun, Sen Su, Jing Shao
arxiv.org/abs/2510.11246

Collaborative Shadows: Distributed Backdoor Attacks in LLM-Based Multi-Agent Systems
LLM-based multi-agent systems (MAS) demonstrate increasing integration into next-generation applications, but their safety in backdoor attacks remains largely underexplored. However, existing research has focused exclusively on single-agent backdoor attacks, overlooking the novel attack surfaces introduced by agent collaboration in MAS. To bridge this gap, we present the first Distributed Backdoor Attack tailored to MAS. We decompose the backdoor into multiple distributed attack primitives that…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:15:28

MPCitH-based Signatures from Restricted Decoding Problems
Michele Battagliola, Sebastian Bitzer, Antonia Wachter-Zeh, Violetta Weger
arxiv.org/abs/2510.11224

MPCitH-based Signatures from Restricted Decoding Problems
Threshold-Computation-in-the-Head (TCitH) and VOLE-in-the-Head (VOLEitH), two recent developments of the MPC-in-the-Head (MPCitH) paradigm, have significantly improved the performance of digital signature schemes in this framework. In this note, we embed the restricted decoding problem within these frameworks. We propose a structurally simple modeling that achieves competitive signature sizes. Specifically, by instantiating the restricted decoding problem with the same hardness assumption und…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:14:08

TraceAegis: Securing LLM-Based Agents via Hierarchical and Behavioral Anomaly Detection
Jiahao Liu, Bonan Ruan, Xianglin Yang, Zhiwei Lin, Yan Liu, Yang Wang, Tao Wei, Zhenkai Liang
arxiv.org/abs/2510.11203

TraceAegis: Securing LLM-Based Agents via Hierarchical and Behavioral Anomaly Detection
LLM-based agents have demonstrated promising adaptability in real-world applications. However, these agents remain vulnerable to a wide range of attacks, such as tool poisoning and malicious instructions, that compromise their execution flow and can lead to serious consequences like data breaches and financial loss. Existing studies typically attempt to mitigate such anomalies by predefining specific rules and enforcing them at runtime to enhance safety. Yet, designing comprehensive rules is di…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:13:08

RAG-Pull: Imperceptible Attacks on RAG Systems for Code Generation
Vasilije Stambolic, Aritra Dhar, Lukas Cavigelli
arxiv.org/abs/2510.11195

RAG-Pull: Imperceptible Attacks on RAG Systems for Code Generation
Retrieval-Augmented Generation (RAG) increases the reliability and trustworthiness of the LLM response and reduces hallucination by eliminating the need for model retraining. It does so by adding external data into the LLM's context. We develop a new class of black-box attack, RAG-Pull, that inserts hidden UTF characters into queries or external code repositories, redirecting retrieval toward malicious code, thereby breaking the models' safety alignment. We observe that query and code perturbat…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:12:18

CoSPED: Consistent Soft Prompt Targeted Data Extraction and Defense
Yang Zhuochen, Fok Kar Wai, Thing Vrizlynn
arxiv.org/abs/2510.11137 arx…

CoSPED: Consistent Soft Prompt Targeted Data Extraction and Defense
Large language models have gained widespread attention recently, but their potential security vulnerabilities, especially privacy leakage, are also becoming apparent. To test and evaluate for data extraction risks in LLM, we proposed CoSPED, short for Consistent Soft Prompt targeted data Extraction and Defense. We introduce several innovative components, including Dynamic Loss, Additive Loss, Common Loss, and Self Consistency Decoding Strategy, and tested to enhance the consistency of the soft …

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:10:18

N-output Mechanism: Estimating Statistical Information from Numerical Data under Local Differential Privacy
Incheol Baek, Yon Dohn Chung
arxiv.org/abs/2510.11116

N-output Mechanism: Estimating Statistical Information from Numerical Data under Local Differential Privacy
Local Differential Privacy (LDP) addresses significant privacy concerns in sensitive data collection. In this work, we focus on numerical data collection under LDP, targeting a significant gap in the literature: existing LDP mechanisms are optimized for either a very small ($|Ω| \in \{2, 3\}$) or infinite output spaces. However, no generalized method for constructing an optimal mechanism for an arbitrary output size $N$ exists. To fill this gap, we propose the \textbf{N-output mechan…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:09:08

Stabilizing the Staking Rate, Dynamically Distributed Inflation and Delay Induced Oscillations
Carlo Brunetta, Amit Chaudhary, Stefano Galatolo, Massimiliano Sala
arxiv.org/abs/2510.11065

Stabilizing the Staking Rate, Dynamically Distributed Inflation and Delay Induced Oscillations
Dynamically distributed inflation is a common mechanism used to guide a blockchain's staking rate towards a desired equilibrium between network security and token liquidity. However, the high sensitivity of the annual percentage yield to changes in the staking rate, coupled with the inherent feedback delays in staker responses, can induce undesirable oscillations around this equilibrium. This paper investigates this instability phenomenon. We analyze the dynamics of inflation-based reward s…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:06:48

Secret-Protected Evolution for Differentially Private Synthetic Text Generation
Tianze Wang, Zhaoyu Chen, Jian Du, Yingtai Xiao, Linjun Zhang, Qiang Yan
arxiv.org/abs/2510.10990

Secret-Protected Evolution for Differentially Private Synthetic Text Generation
Text data has become extremely valuable on large language models (LLMs) and even lead to general artificial intelligence (AGI). A lot of high-quality text in the real world is private and cannot be freely used due to privacy concerns. Therefore, differentially private (DP) synthetic text generation has been proposed, aiming to produce high-utility synthetic data while protecting sensitive information. However, existing DP synthetic text generation imposes uniform guarantees that often overprote…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:05:38

DITTO: A Spoofing Attack Framework on Watermarked LLMs via Knowledge Distillation
Hyeseon Ahn, Shinwoo Park, Yo-Sub Han
arxiv.org/abs/2510.10987

DITTO: A Spoofing Attack Framework on Watermarked LLMs via Knowledge Distillation
The promise of LLM watermarking rests on a core assumption that a specific watermark proves authorship by a specific model. We demonstrate that this assumption is dangerously flawed. We introduce the threat of watermark spoofing, a sophisticated attack that allows a malicious model to generate text containing the authentic-looking watermark of a trusted, victim model. This enables the seamless misattribution of harmful content, such as disinformation, to reputable sources. The key to our attack…

@arXiv_csCR_bot@mastoxiv.page
2025-10-14 12:04:28

TabVLA: Targeted Backdoor Attacks on Vision-Language-Action Models
Zonghuan Xu, Xiang Zheng, Xingjun Ma, Yu-Gang Jiang
arxiv.org/abs/2510.10932

TabVLA: Targeted Backdoor Attacks on Vision-Language-Action Models
With the growing deployment of Vision-Language-Action (VLA) models in real-world embodied AI systems, their increasing vulnerability to backdoor attacks poses a serious safety threat. A backdoored VLA agent can be covertly triggered by a pre-injected backdoor to execute adversarial actions, potentially causing system failures or even physical harm. Although backdoor attacks on VLA models have been explored, prior work has focused only on untargeted attacks, leaving the more practically threaten…