Papers for

model governance teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Framework measures response variability and risk in large language models

A Distributional Framework for Behavioral Characterization and Risk Management of Large Language Models

Abstract: Large language models are stochastic by nature. Repeated executions under the same experimental conditions may produce different responses. We propose a framework that can be applied across the model lifecycle, from task design to residual behavior and continuous monitoring. The method consists of building an empirical response distribution estimated through Monte Carlo simulation. Semantic interventions are introduced to study how these distributions change. We illustrate the framework with two different tasks, borrower classification and music album recommendation, and quantify the resulting distributional changes. The framework is linked to model risk management and governance by measuring robustness and control effectiveness, and characterizing residual behavior. Since it relies only on observed responses, it can also be applied to black-box models accessed through commercial APIs.

Thu 8 OctMachine Learning
The gist
Large language models can give different answers each time you ask the same question because they work in a random way. The authors propose a way to study all possible answers by running the model many times and seeing how the answers change. They also test what happens when they try to steer or change the model’s behavior in specific ways. This helps people understand how reliable the model is, how well it can be controlled, and what unexpected behaviors might remain. Their approach works even if you only have access to the model’s outputs, like when using online services.
Open → 2610.12718v1