Predictive Credit: Measuring What Scientific Explanations Add to Experimental Forecasts Paper • 2610.00314 • Published Sep 29 • 9
ControlScope: Workflow Revision and Reliability in LLM Agents Paper • 2609.34313 • Published Sep 28 • 3
WhatWorkedBench: Benchmarking Experimental Understanding in AI Agents Paper • 2609.27490 • Published Sep 23 • 14