Fourier circuits in neural networks: Unlocking the potential of large language models in mathematical reasoning and modular arithmetic
In the evolving landscape of machine learning, a pivotal challenge lies in deciphering the
internal representations harnessed by neural networks and Transformers. Building on recent …
internal representations harnessed by neural networks and Transformers. Building on recent …
The evolution of statistical induction heads: In-context learning markov chains
Large language models have the ability to generate text that mimics patterns in their inputs.
We introduce a simple Markov Chain sequence modeling task in order to study how this in …
We introduce a simple Markov Chain sequence modeling task in order to study how this in …
Opening the black box of large language models: Two views on holistic interpretability
As large language models (LLMs) grow more powerful, concerns around potential harms
like toxicity, unfairness, and hallucination threaten user trust. Ensuring beneficial alignment …
like toxicity, unfairness, and hallucination threaten user trust. Ensuring beneficial alignment …
Out-of-distribution generalization via composition: a lens through induction heads in transformers
J Song, Z Xu, Y Zhong - arXiv preprint arXiv:2408.09503, 2024 - arxiv.org
Large language models (LLMs) such as GPT-4 sometimes appear to be creative, solving
novel tasks often with a few demonstrations in the prompt. These tasks require the models to …
novel tasks often with a few demonstrations in the prompt. These tasks require the models to …
Why do you grok? a theoretical analysis of grokking modular addition
We present a theoretical explanation of the``grokking''phenomenon, where a model
generalizes long after overfitting, for the originally-studied problem of modular addition. First …
generalizes long after overfitting, for the originally-studied problem of modular addition. First …
Interpreting grokked transformers in complex modular arithmetic
Grokking has been actively explored to reveal the mystery of delayed generalization.
Identifying interpretable algorithms inside the grokked models is a suggestive hint to …
Identifying interpretable algorithms inside the grokked models is a suggestive hint to …
Deep Learning Through A Telescoping Lens: A Simple Model Provides Empirical Insights On Grokking, Gradient Boosting & Beyond
Deep learning sometimes appears to work in unexpected ways. In pursuit of a deeper
understanding of its surprising behaviors, we investigate the utility of a simple yet accurate …
understanding of its surprising behaviors, we investigate the utility of a simple yet accurate …
Approaching deep learning through the spectral dynamics of weights
We propose an empirical approach centered on the spectral dynamics of weights--the
behavior of singular values and vectors during optimization--to unify and clarify several …
behavior of singular values and vectors during optimization--to unify and clarify several …
Emergence in non-neural models: grokking modular arithmetic via average gradient outer product
Neural networks trained to solve modular arithmetic tasks exhibit grokking, a phenomenon
where the test accuracy starts improving long after the model achieves 100% training …
where the test accuracy starts improving long after the model achieves 100% training …
Gradient descent induces alignment between weights and the empirical NTK for deep non-linear networks
Understanding the mechanisms through which neural networks extract statistics from input-
label pairs is one of the most important unsolved problems in supervised learning. Prior …
label pairs is one of the most important unsolved problems in supervised learning. Prior …