Zobrazeno 1 - 10
of 2 013
pro vyhledávání: '"Dejean, P."'
Autor:
Déjean, Hervé
In this paper, we investigate how efficiently large language models (LLM) can be trained to check whether an answer is already stored in their parametric memory. We distill an LLM-as-a-judge to compute the IK (I Know) score. We found that this method
Externí odkaz:
http://arxiv.org/abs/2412.11536
Autor:
Gatti, G., Tancogne-Dejean, N., Hübener, H., De Giovannini, U., Dai, J., Polishchuk, S., Bugnon, Ph., Frassetto, F., Poletto, L., Chergui, M., Grioni, M., Rubio, A., Puppin, M., Crepaldi, A.
Chirality in tellurium derives from a Peierls distortion driven by strong electron-phonon coupling, making this material a unique candidate for observing a light-induced topological phase transition. By using time- and angle-resolved photoelectron sp
Externí odkaz:
http://arxiv.org/abs/2411.13954
A beyond electric-dipole light-matter theory is needed to describe emerging X-ray and THz applications for characterization and control of quantum materials but inaccessible as nondipole lattice-aperiodic terms impede on the use of Bloch's theorem. T
Externí odkaz:
http://arxiv.org/abs/2410.18547
Semantic navigation enables robots to understand their environments beyond basic geometry, allowing them to reason about objects, their functions, and their interrelationships. In semantic robotic navigation, creating accurate and semantically enrich
Externí odkaz:
http://arxiv.org/abs/2410.14851
Publikováno v:
WSDM 2025
Retrieval-Augmented Generation (RAG) allows overcoming the limited knowledge of LLMs by extending the input with external information. As a consequence, the contextual inputs to the model become much longer which slows down decoding time directly tra
Externí odkaz:
http://arxiv.org/abs/2407.09252
Autor:
Rau, David, Déjean, Hervé, Chirkova, Nadezhda, Formal, Thibault, Wang, Shuai, Nikoulina, Vassilina, Clinchant, Stéphane
Retrieval-Augmented Generation allows to enhance Large Language Models with external knowledge. In response to the recent popularity of generative LLMs, many RAG approaches have been proposed, which involve an intricate number of different configurat
Externí odkaz:
http://arxiv.org/abs/2407.01102
Autor:
Chirkova, Nadezhda, Rau, David, Déjean, Hervé, Formal, Thibault, Clinchant, Stéphane, Nikoulina, Vassilina
Retrieval-augmented generation (RAG) has recently emerged as a promising solution for incorporating up-to-date or domain-specific knowledge into large language models (LLMs) and improving LLM factuality, but is predominantly studied in English-only s
Externí odkaz:
http://arxiv.org/abs/2407.01463
Nonlinear photomagnetization is a process by which an oscillating electric field can induce a static magnetization. We show that all 32 crystallographic point groups admit spin polarization to second order using circularly polarized electric fields (
Externí odkaz:
http://arxiv.org/abs/2406.14748
Autor:
Xu, Qiaoling, Tancogne-Dejean, Nicolas, Boström, Emil Viñas, Kennes, Dante M., Claassen, Martin, Rubio, Angel, Xian, Lede
Due to the large-period superlattices emerging in moir\'e two-dimensional (2D) materials, electronic states in such systems exhibit low energy flat bands that can be used to simulate strongly correlated physics in a highly tunable setup. While many i
Externí odkaz:
http://arxiv.org/abs/2406.05626
The late interaction paradigm introduced with ColBERT stands out in the neural Information Retrieval space, offering a compelling effectiveness-efficiency trade-off across many benchmarks. Efficient late interaction retrieval is based on an optimized
Externí odkaz:
http://arxiv.org/abs/2404.13950