Search Results for author:
Found 10 papers, 5 papers with code
Auditing language models for hidden objectives
First, we conduct a blind auditing game where four teams, unaware of the model's hidden objective or training, investigate it for concerning behaviors and their causes.
Alignment faking in large language models
Explaining this gap, in almost all cases where the model complies with a harmful query from a free user, we observe explicit alignment-faking reasoning, with the model stating it is strategically answering harmful queries in training to preserve its preferred harmlessness behavior out of training.
Connecting the Dots: LLMs can Infer and Verbalize Latent Structure from Disparate Training Data
As a step towards answering this question, we study inductive out-of-context reasoning (OOCR), a type of generalization in which LLMs infer latent information from evidence distributed across training documents and apply it to downstream tasks without in-context learning.
Modeling evidential cooperation in large worlds
I discuss gains from trade given uncertain beliefs about other agents and analyze how these gains decrease in several toy examples as the belief in another agent decreases.
Incentivizing honest performative predictions with proper scoring rules
We show that, for binary predictions, if the influence of the expert's prediction on outcomes is bounded, it is possible to define scoring rules under which optimal reports are arbitrarily close to fixed points.
Conditioning Predictive Models: Risks and Strategies
Our intention is to provide a definitive reference on what it would take to safely make use of generative/predictive models in the absence of a solution to the Eliciting Latent Knowledge problem.
Path Independent Equilibrium Models Can Better Exploit Test-Time Computation
Designing networks capable of attaining better performance with an increased inference budget is important to facilitate generalization to harder problem instances.
COLA: Consistent Learning with Opponent-Learning Awareness
Finally, in an empirical evaluation on a set of general-sum games, we find that COLA finds prosocial solutions and that it converges under a wider range of learning rates than HOLA and LOLA.
Normative Disagreement as a Challenge for Cooperative AI
However, the mixed-motive environments typically studied have a single cooperative outcome on which all agents can agree.
A New Formalism, Method and Open Issues for Zero-Shot Coordination
We introduce an extension of the algorithm, other-play with tie-breaking, and prove that it is optimal in the LFC problem and an equilibrium in the LFC game.
