activity
20242026
collaborators
Showing cs.LGShow all

5 papers · 1 filter

cs.LG2026

BAITBENCH: Measuring Agent Reward Hacking with Optional Shortcuts Planted in ML Tasks

Pradyumna Shyama Prasad, Meiri Anto, Leon Eshuijs +3

LLM agents are increasingly used to run autonomous ML experiments, iterating on target metrics with little human oversight. Prior work has documented reward hacking in these enviro…

cs.LG2026

Safety Training Modulates Harmful Misalignment Under On-Policy RL, But Direction Depends on Environment Design

Leon Eshuijs, Shihan Wang, Antske Fokkens

Specification gaming under Reinforcement Learning (RL) is known to cause LLMs to develop sycophantic, manipulative, or deceptive behavior, yet the conditions under which this occur…

cs.LG2025

Short-circuiting Shortcuts: Mechanistic Investigation of Shortcuts in Text Classification

Leon Eshuijs, Shihan Wang, Antske Fokkens

Reliance on spurious correlations (shortcuts) has been shown to underlie many of the successes of language models. Previous work focused on identifying the input elements that impa…

cs.LG2025

But what is your honest answer? Aiding LLM-judges with honest alternatives using steering vectors

Leon Eshuijs, Archie Chaudhury, Alan McBeth +1

LLM-as-a-judge is widely used as a scalable substitute for human evaluation, yet current approaches rely on black-box access and struggle to detect subtle dishonesty, such as sycop…

cs.LG2024

Balancing the Scales: Reinforcement Learning for Fair Classification

Leon Eshuijs, Shihan Wang, Antske Fokkens

Fairness in classification tasks has traditionally focused on bias removal from neural representations, but recent trends favor algorithmic methods that embed fairness into the tra…