Results 121 to 130 of about 5,833,459 (242)
Fine-tuning large language models (LLMs) for specific tasks introduces privacy risks, as models may inadvertently memorise and leak sensitive training data.
Passerat-Palmbach, Jonathan +2 more
core +1 more source
Photo‐Degradable Polyester Networks and Multi‐Photon Printed Objects Based on Cyclic Ketene Acetals
Advanced Functional Materials, EarlyView.The current work introduces a photoreversible polyester network derived from radical ring‐opening polymerization of cyclic ketene acetals. It combines photoreversible cross‐linking with initiator‐free multi‐photon printing. Reversible network formation, tunable mechanical properties, and selective degradation highlight its potential as a versatile ...
Till Meissner +5 more
wiley +1 more source
Electron‐Deficient Linkers Enhance H2O2 Electrosynthesis in Covalent Organic Frameworks
Advanced Functional Materials, EarlyView.Linker π‐conjugation modulates the electronic state of the Ni–N4 centers and their interaction with the key *OOH intermediate while preserving the primary coordination motif. NiPc‐PTDA delivers an H2O2 selectivity of 92% and a production rate of 34.8 mol gcat−1 h−1, highlighting linker electronic properties as a molecular design parameter for NiPc ...
Houting Xie +7 more
wiley +1 more source
Parameter-Efficient Fine-Tuning With Adapters
In the arena of language model fine-tuning, the traditional approaches, such as Domain-Adaptive Pretraining (DAPT) and Task-Adaptive Pretraining (TAPT), although effective, but computational intensive.
Pang, Yuan, Chen, Keyu, Yang, Zi
core
Architecture‐Driven Functional Coupling in Vertically Aligned Nanocomposites
Advanced Functional Materials, EarlyView.Vertically aligned nanocomposites define a growth‐engineered architecture in which vertical interfaces, strain fields, defect pathways, and phase connectivity are created simultaneously. This review shows how these architectural features couple ferroic, optical, ionic, electrochemical, and device responses, establishing design rules and open challenges
Md Shatil Islam‐Shanto +4 more
wiley +1 more source
Advanced Functional Materials, EarlyView.An Alhagi maurorum inspired hierarchical composite membrane (NFCu) was developed by integrating heavy metal remediation and solar‐driven interfacial evaporation in a single platform. Beyond the photothermal effect, a localized photocatalytic process is realized on the CuO surface, deepening the mechanistic understanding of light–matter–water ...
Yuqiong He +9 more
wiley +1 more source
Light-PEFT: Lightening Parameter-Efficient Fine-Tuning via Early Pruning [PDF]
Parameter-efficient fine-tuning (PEFT) has emerged as the predominant technique for fine-tuning in the era of large language models. However, existing PEFT methods still have inadequate training efficiency.
Gu, Naibin +5 more
core
Covalent Functionalization of 2D Semiconductors: A Roadmap to Advanced Electronic Devices
Advanced Functional Materials, EarlyView.This Review presents recent advances in the covalent functionalization strategies for two‐dimensional semiconductors and their implementation in modern technologies. Layered materials are modified through diverse molecular chemistries (e.g., thiols, diazonium salts, alkyl halides, and electron‐deficient species) to tailor their surface properties ...
Ramiro Quirós‐Ovies +2 more
wiley +1 more source
RoSA: Accurate parameter-efficient fine-tuning via robust adaptation
We investigate parameter-efficient fine-tuning (PEFT) methods that can provide good accuracy under limited computational and memory budgets in the context of large language models (LLMs).
Crncevic, Elvir +3 more
core +3 more sources
Strong Baselines for Parameter-Efficient Few-Shot Fine-Tuning
Few-shot classification (FSC) entails learning novel classes given only a few examples per class after a pre-training (or meta-training) phase on a set of base classes. Recent works have shown that simply fine-tuning a pre-trained Vision Transformer (ViT)
Basu, Samyadeep +3 more
core +1 more source