8 citations · 9 across the 7 of their papers we have counts for
9 papers
Mitigating Social Sycophancy via Pluralistic Preference Optimization
Stephane Hatgis-Kessell, Myra Cheng, Xiaoxuan Hou +4
Personal advice, including relationship advice, now ranks among the most common uses of generative AI. But language models (LMs) exhibit sycophancy: they affirm users much more oft…
Specifying Reward Functions for RL Without Environment Sampling
Stephane Hatgis-Kessell, W. Bradley Knox, Emma Brunskill
Enabling human stakeholders to specify reward functions that lead to their desired outcomes is a key challenge in deploying reinforcement learning agents. Preference-based methods…
Estimating time spent on work tasks
Stephane Hatgis-Kessell, Tomás Aguirre, Alexander Wan +1
The task-based framework in economics models occupations as bundles of tasks. It is the standard lens for understanding how technology affects work: a new technology changes the co…
Economic Evaluations of Language Models
Alexander Wan, Stephane Hatgis-Kessell, Tomás Aguirre +2
Language models perform economically valuable work, yet they are not currently assessed for how well they perform every economically valuable task. We introduce EconEvals as an ope…
When are LLMs Sufficient Policy Optimizers for Sequential RL Tasks?
Stephane Hatgis-Kessell, Emma Brunskill
We study when large language models (LLMs) can serve as effective black-box policy optimizers for reinforcement learning (RL) tasks, i.e., when can we replace classical RL algorith…
Repairing Reward Functions with Feedback to Mitigate Reward Hacking
Stephane Hatgis-Kessell, Logan Mondal Bhamidipaty, Emma Brunskill
Human-designed reward functions for reinforcement learning (RL) agents are frequently misaligned with the humans' true, unobservable objectives, and thus act only as proxies. Optim…
Influencing Humans to Conform to Preference Models for RLHF
Stephane Hatgis-Kessell, W. Bradley Knox, Serena Booth +1
Designing a reinforcement learning from human feedback (RLHF) algorithm to approximate a human's unobservable reward function requires assuming, implicitly or explicitly, a model o…
Learning Optimal Advantage from Preferences and Mistaking it for Reward
W. Bradley Knox, Stephane Hatgis-Kessell, Sigurdur Orn Adalgeirsson +4
We consider algorithms for learning reward functions from human preferences over pairs of trajectory segments, as used in reinforcement learning from human feedback (RLHF). Most re…
Models of human preference for learning reward functions
W. Bradley Knox, Stephane Hatgis-Kessell, Serena Booth +3
The utility of reinforcement learning is limited by the alignment of reward functions with the interests of human stakeholders. One promising method for alignment is to learn the r…