Author name cluster

Emily Reif

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

1 author row

NeurIPS Conference 2024 Conference Paper

Who's asking? User personas and the mechanics of latent misalignment

Asma Ghandeharioun
Ann Yuan
Marius Guerard
Emily Reif
Michael A. Lepori
Lucas Dixon

Studies show that safety-tuned models may nevertheless divulge harmful information. In this work, we show that whether they do so depends significantly on who they are talking to, which we refer to as user persona. In fact, we find manipulating user persona to be more effective for eliciting harmful content than certain more direct attempts to control model refusal. We study both natural language prompting and activation steering as intervention methods and show that activation steering is significantly more effective at bypassing safety filters. We shed light on the mechanics of this phenomenon by showing that even when model generations are safe, harmful content can persist in hidden representations and can be extracted by decoding from earlier layers. We also show we can predict a persona’s effect on refusal given only the geometry of its steering vector. Finally, we show that certain user personas induce the model to form more charitable interpretations of otherwise dangerous queries.

PDF Details DOI

JMLR Journal 2023 Journal Article

PaLM: Scaling Language Modeling with Pathways

Aakanksha Chowdhery
Sharan Narang
Jacob Devlin
Maarten Bosma
Gaurav Mishra
Adam Roberts
Paul Barham
Hyung Won Chung

Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model (PaLM). We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning results on hundreds of language understanding and generation benchmarks. On a number of these tasks, PaLM 540B achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark. A significant number of BIG-bench tasks showed discontinuous improvements from model scale, meaning that performance steeply increased as we scaled to our largest model. PaLM also has strong capabilities in multilingual tasks and source code generation, which we demonstrate on a wide array of benchmarks. We additionally provide a comprehensive analysis on bias and toxicity, and study the extent of training data memorization with respect to model scale. Finally, we discuss the ethical considerations related to large language models and discuss potential mitigation strategies. [abs] [ pdf ][ bib ] &copy JMLR 2023. ( edit, beta )

PDF Details

NeurIPS Conference 2020 Conference Paper

Evaluating Attribution for Graph Neural Networks

Benjamin Sanchez-Lengeling
Jennifer Wei
Brian Lee
Emily Reif
Peter Wang
Wesley Qian
Kevin McCloskey
Lucy Colwell

Interpretability of machine learning models is critical to scientific understanding, AI safety, as well as debugging. Attribution is one approach to interpretability, which highlights input dimensions that are influential to a neural network’s prediction. Evaluation of these methods is largely qualitative for image and text models, because acquiring ground truth attributions requires expensive and unreliable human judgment. Attribution has been little studied for graph neural networks (GNNs), a model class of growing importance that makes predictions on arbitrarily-sized graphs. In this work we adapt commonly-used attribution methods for GNNs and quantitatively evaluate them using computable ground-truths that are objective and challenging to learn. We make concrete recommendations for which attribution methods to use, and provide the data and code for our benchmarking suite. Rigorous and open source benchmarking of attribution methods in graphs could enable new methods development and broader use of attribution in real-world ML tasks.

PDF Details

NeurIPS Conference 2019 Conference Paper

Visualizing and Measuring the Geometry of BERT

Emily Reif
Ann Yuan
Martin Wattenberg
Fernanda Viegas
Andy Coenen
Adam Pearce
Been Kim

Transformer architectures show significant promise for natural language processing. Given that a single pretrained model can be fine-tuned to perform well on many different tasks, these networks appear to extract generally useful linguistic features. A natural question is how such networks represent this information internally. This paper describes qualitative and quantitative investigations of one particularly effective model, BERT. At a high level, linguistic features seem to be represented in separate semantic and syntactic subspaces. We find evidence of a fine-grained geometric representation of word senses. We also present empirical descriptions of syntactic representations in both attention matrices and individual word embeddings, as well as a mathematical argument to explain the geometry of these representations.

PDF Details

AAAI Conference 2018 Conference Paper

Bandit-Based Solar Panel Control

David Abel
Edward Williams
Stephen Brawner
Emily Reif
Michael Littman

Solar panels sustainably harvest energy from the sun. To improve performance, panels are often equipped with a tracking mechanism that computes the sun’s position in the sky throughout the day. Based on the tracker’s estimate of the sun’s location, a controller orients the panel to minimize the angle of incidence between solar radiant energy and the photovoltaic cells on the surface of the panel, increasing total energy harvested. Prior work has developed efﬁcient tracking algorithms that accurately compute the sun’s location to facilitate solar tracking and control. However, always pointing a panel directly at the sun does not account for diffuse irradiance in the sky, reﬂected irradiance from the ground and surrounding surfaces, power required to reorient the panel, shading effects from neighboring panels and foliage, or changing weather conditions (such as clouds), all of which are contributing factors to the total energy harvested by a ﬂeet of solar panels. In this work, we show that a bandit-based approach can increase the total energy harvested by solar panels by learning to dynamically account for such other factors. Our contribution is threefold: (1) the development of a test bed based on typical solar and irradiance models for experimenting with solar panel control using a variety of learning methods, (2) simulated validation that bandit algorithms can effectively learn to control solar panels, and (3) the design and construction of an intelligent solar panel prototype that learns to angle itself using bandit algorithms.

PDF Details

RLDM Conference 2017 Conference Abstract

Improving Solar Panel Efficiency Using Reinforcement Learning

David Abel
Emily Reif
Michael Littman

Solar panels sustainably harvest energy from the sun. To improve performance, panels are often equipped with a tracking mechanism that computes the sun’s position in the sky throughout the day. Based on the tracker’s estimate of the sun’s location, a controller orients the panel to minimize the angle of inci- dence between solar radiant energy and the photovoltaic cells on the surface of the panel, increasing total energy harvested. Prior work has developed efficient tracking algorithms that accurately compute the sun’s location to facilitate solar tracking and control. However, always pointing a panel directly at the sun does not account for diffuse irradiance in the sky, reflected irradiance from the ground and surrounding surfaces, or changing weather conditions (such as cloud coverage), all of which are contributing factors to the total energy harvested by a solar panel. In this work, we show that a reinforcement learning (RL) approach can increase the total energy harvested by solar panels by learning to dynamically account for such other factors. We advocate for the use of RL for solar panel control due to its effectiveness, negligible cost, and versatility. Our contribution is twofold: (1) an adaption of typical RL algorithms to the task of improving solar panel performance, and (2) an experimental validation in simulation based on typical solar and irradiance models for experimenting with solar panel control. We evaluate the utility of various RL approaches compared to an idealized controller, an efficient state-of-the-art direct tracking algorithm, and a fixed panel in our simulated environment. We experiment across different time scales, in different places on earth, and with dramati- cally different percepts (sun coordinates and raw images of the sky with and without clouds), consistently demonstrating that simple RL algorithms improve over existing baselines.

PDF Details