Shortened llama: A simple depth pruning for large language models
Structured pruning of modern large language models (LLMs) has emerged as a way of
decreasing their high computational needs. Width pruning reduces the size of projection …
decreasing their high computational needs. Width pruning reduces the size of projection …
Opportunities and Challenges in Transformer Neural Networks for Battery State Estimation: Charge, Health, Lifetime, and Safety
Battery technology plays a crucial role across various sectors, powering devices from
smartphones to electric vehicles and supporting grid-scale energy storage. To ensure their …
smartphones to electric vehicles and supporting grid-scale energy storage. To ensure their …
Elms: Elasticized large language models on mobile devices
On-device Large Language Models (LLMs) are revolutionizing mobile AI, enabling
applications such as UI automation while addressing privacy concerns. Currently, the …
applications such as UI automation while addressing privacy concerns. Currently, the …
Towards More Trustworthy and Interpretable LLMs for Code through Syntax-Grounded Explanations
DN Palacio, D Rodriguez-Cardenas, A Velasco… - arxiv preprint arxiv …, 2024 - arxiv.org
Trustworthiness and interpretability are inextricably linked concepts for LLMs. The more
interpretable an LLM is, the more trustworthy it becomes. However, current techniques for …
interpretable an LLM is, the more trustworthy it becomes. However, current techniques for …
The openelm library: Leveraging progress in language models for novel evolutionary algorithms
Abstract In recent years, Large Language Models (LLMs) have rapidly progressed in their
capabilities in natural language processing (NLP) tasks, which have interestingly grown in …
capabilities in natural language processing (NLP) tasks, which have interestingly grown in …
Evolutions of semantic consistency in research topic via contextualized word embedding
Topic evolution has been studied extensively in the field of the science of science. This study
first analyzes topic evolution pattern from topics' semantic consistency in the semantic vector …
first analyzes topic evolution pattern from topics' semantic consistency in the semantic vector …
LoRA : Multi-Scale Low-Rank Approximations for Fine-Tuning Large Language Models
JC Zhang, YJ **ong, HX Qiu, DH Zhu… - arxiv preprint arxiv …, 2024 - arxiv.org
Fine-tuning large language models (LLMs) with high parameter efficiency for downstream
tasks has become a new paradigm. Low-Rank Adaptation (LoRA) significantly reduces the …
tasks has become a new paradigm. Low-Rank Adaptation (LoRA) significantly reduces the …
Parameter-Efficient Fine-Tuning of Large Language Models via Deconvolution in Subspace
JC Zhang, YJ **ong, CM **a, DH Zhu… - Proceedings of the 31st …, 2025 - aclanthology.org
This paper proposes a novel parameter-efficient fine-tuning method that combines the
knowledge completion capability of deconvolution with the subspace learning ability …
knowledge completion capability of deconvolution with the subspace learning ability …
Bilateral Cross-Modal Fusion Network for Multimodal Whole-Body Tumor Segmentation
3D whole-body tumor segmentation plays a critical role in cancer treatment. In relation to
this, multimodal fluorodeoxyglucose (FDG) positron emission tomography/computed …
this, multimodal fluorodeoxyglucose (FDG) positron emission tomography/computed …
Presidential Communication and Its Impact on the Mexican Stock Market: Evidence Using a Sentiment Analysis Approach
JA Cazares Aguilar… - Latin American Business …, 2024 - Taylor & Francis
This research paper addresses the impact of the positive or negative polarity of the
presidential messages contained in the stenographic version of the daily conferences …
presidential messages contained in the stenographic version of the daily conferences …