Zobrazeno 1 - 10
of 11 426
pro vyhledávání: '"Berezin, A."'
We introduce a novel family of adversarial attacks that exploit the inability of language models to interpret ASCII art. To evaluate these attacks, we propose the ToxASCII benchmark and develop two custom ASCII art fonts: one leveraging special token
Externí odkaz:
http://arxiv.org/abs/2409.18708
This paper addresses the challenge of neural state estimation in power distribution systems. We identified a research gap in the current state of the art, which lies in the inability of models to adapt to changes in the power grid, such as loss of se
Externí odkaz:
http://arxiv.org/abs/2408.05787
Autor:
Berezin, A. V., Fedotov, A. M.
Publikováno v:
Phys. Rev. D 110, 016009 (2024)
We consider a setup to detect stimulated photon-photon scattering using high-power lasers. Signal photons are emitted from an overlap of the incoming intense laser pulses focused in vacuum from three sides. We derive and justify a general approximate
Externí odkaz:
http://arxiv.org/abs/2405.00151
Autonomous and learning systems based on Deep Reinforcement Learning have firmly established themselves as a foundation for approaches to creating resilient and efficient Cyber-Physical Energy Systems. However, most current approaches suffer from two
Externí odkaz:
http://arxiv.org/abs/2404.01794
Autor:
Berezin, Victor, Ivanova, Inna
Publikováno v:
Universe 2024, 10(3), 147
The action of an ideal fluid in Euler variables with a variable number of particles is used for the phenomenological description of the processes of particle creation in strong external fields. It has been demonstrated that the conformal invariance o
Externí odkaz:
http://arxiv.org/abs/2403.14031
Short review of the Weyl geometry is given. To describe the phenomenological particle creation we suggest the modified perfect fluid model taking into account the back reaction on the geometry of both the already created particles and the very proces
Externí odkaz:
http://arxiv.org/abs/2401.06235
Autor:
Berezin, Victor, Ivanova, Inna
Using a model for an ideal fluid with a variable number of particles, a phenomenological description of the processes of particle production in strong external fields is investigated. The conformal invariance of the creation law is shown, which impos
Externí odkaz:
http://arxiv.org/abs/2312.07237
We introduce a simple yet efficient sentence-level attack on black-box toxicity detector models. By adding several positive words or sentences to the end of a hateful message, we are able to change the prediction of a neural network and pass the toxi
Externí odkaz:
http://arxiv.org/abs/2310.13099
The fundamental problem in toxicity detection task lies in the fact that the toxicity is ill-defined. This causes us to rely on subjective and vague data in models' training, which results in non-robust and non-accurate results: garbage in - garbage
Externí odkaz:
http://arxiv.org/abs/2310.02357
Autor:
Berezin, Sergey
We consider the hard-edge scaling of the Mittag-Leffler ensemble confined to a fixed disk inside the droplet. Our primary emphasis is on fluctuations of rotationally-invariant additive statistics that depend on the radius and thus give rise to radius
Externí odkaz:
http://arxiv.org/abs/2308.12658