AutoSaddler: Automatic Harness Optimization with Durable Updates from Agent Execution Traces Paper • 2608.23041 • Published 10 days ago • 64
Dayhoff Atlas Collection The models and datasets that comprise the Dayhoff Atlas • 11 items • Updated 8 days ago • 13
One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflows Paper • 2608.19741 • Published 14 days ago • 12