Jacob Andreas

Shoot First, Ask Questions Later? Building Rational Agents that Explore and Act Like People

Language and Experience: A Computational Model of Social Learning in Complex Tasks

Log-Augmented Generation: Scaling Test-Time Reasoning with Reusable Computation

Sequential Parallel Duality in Prefix Scannable Models

CONCUR: A Framework for Continual Constrained and Unconstrained Routing

Beyond Binary Rewards: Training LMs to Reason About Their Uncertainty

Learning To Make MISTAKEs: Modeling Incorrect Student Thinking And Key Errors

OpenEstimate: Evaluating LLMs on Reasoning Under Uncertainty with Real-World Data

Modeling Student Learning with 3.8 Million Program Traces

Code-enabled language models can outperform reasoning models on diverse tasks

Optimizable LLM Planning: A Branch-and-Bound Framework for Complex Tasks

Learning Linear Attention in Polynomial Time

Automated Detection of Visual Attribute Reliance with a Self-Reflective Agent

Eliciting Human Preferences with Language Models

Breakpoint: Stress-testing systems-level reasoning in LLM agents

(How) Do Language Models Track State?

Learning How Hard to Think: Input-Adaptive Allocation of LM Computation

LoRA vs Full Fine-tuning: An Illusion of Equivalence

A Hitchhiker's Guide to Scaling Law Estimation

Self-Steering Language Models

A Hitchhiker's Guide to Scaling Law Estimation

LoRA vs Full Fine-tuning: An Illusion of Equivalence

The Surprising Effectiveness of Test-Time Training for Few-Shot Learning