The illusion of state in state-space models (arxiv.org)

Screenshot of arxiv.org: The illusion of state in state-space models

First screen, captured 2 years ago.

Looks similar

Screenshot of arxiv.org: Mamba: Linear-Time Sequence Modeling with Selective State Spaces

Mamba: Linear-Time Sequence Modeling with Selective State Spaces (arxiv.org)

76 points by anigbrowl 2 years ago | 20 comments

Screenshot of arxiv.org: Matryoshka Representation Learning

Matryoshka Representation Learning (arxiv.org)

71 points by fzliu 2 years ago | 10 comments

Screenshot of arxiv.org: Language Models Represent Space and Time

Language Models Represent Space and Time (arxiv.org)

103 points by birriel 3 years ago | 96 comments

Screenshot of arxiv.org: LoRA vs. Full Fine-Tuning: An Illusion of Equivalence

LoRA vs. Full Fine-Tuning: An Illusion of Equivalence (arxiv.org)

223 points by timbilt 1 year ago | 50 comments

Screenshot of arxiv.org: The universal weight subspace hypothesis

The universal weight subspace hypothesis (arxiv.org)

344 points by lukeplato 10 months ago | 121 comments

arXiv paper proposing that deep neural networks trained on diverse tasks converge to shared low-dimensional weight subspaces, based on spectral analysis of 1100+ models.

neural-network-weights spectral-analysis lora vision-transformers

Screenshot of arxiv.org: On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs

On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs (arxiv.org)

28 points by bilsbie 10 months ago | 38 comments

Looks similar

Mamba: Linear-Time Sequence Modeling with Selective State Spaces(arxiv.org)

Matryoshka Representation Learning(arxiv.org)

Language Models Represent Space and Time(arxiv.org)

LoRA vs. Full Fine-Tuning: An Illusion of Equivalence(arxiv.org)

The universal weight subspace hypothesis(arxiv.org)

On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs(arxiv.org)