The rise and potential of large language model based agents: A survey

Z ** Transformer with Random-Access Reading for Long-Context Understanding
C Yang, Z Yang, N Hua - arxiv preprint arxiv:2405.13216, 2024 - arxiv.org
Long-context modeling presents a significant challenge for transformer-based large
language models (LLMs) due to the quadratic complexity of the self-attention mechanism …

Out-of-distribution generalisation in spoken language understanding

D Porjazovski, A Moisio, M Kurimo - arxiv preprint arxiv:2407.07425, 2024 - arxiv.org
Test data is said to be out-of-distribution (OOD) when it unexpectedly differs from the training
data, a common challenge in real-world use cases of machine learning. Although OOD …

[PDF][PDF] A comprehensive study on LLM agent challenges

P Ingle, M Parab, P Lendave, A Bhanushali, PK Bn - aair-lab.github.io
This paper intricately examines the manifold challenges and inherent issues associated with
Large Language Models (LLMs), both for the models themselves and the human context. It …

Length Generalization with Recursive Neural Networks and Beyond

JR Chowdhury - 2024 - search.proquest.com
Abstract We investigate Recursive Neural Networks (RvNNs) for language processing tasks.
Roughly, from a generalized perspective, RvNNs repeatedly apply some neural function on …