Self-Retrospection Distillation: Turning Post-hoc Experiences into Prior Foresight Paper • 2610.08077 • Published 12 days ago • 196
CoTrace: Data Recipes for Training Terminal Agents with Harness-Model Co-Evolution Paper • 2610.10426 • Published 11 days ago • 12
WhiteMatter: All-to-All Cross-Layer Connections via KV Source Mixing Paper • 2608.18486 • Published 21 days ago • 5
DarwinX: Evolving Agent Harnesses Through Natural Selection Paper • 2608.07545 • Published Jul 31 • 121
AI4AI at Test-Time: Strong-to-Weak Capability Transfer via Harnesses Paper • 2608.12307 • Published Aug 12 • 120
The Illusion of Certainty: Decoupling Capability and Calibration in On-Policy Distillation Paper • 2604.16830 • Published Apr 18 • 15
SkillOrchestra: Learning to Route Agents via Skill Transfer Paper • 2602.19672 • Published Feb 23 • 59
How Predictable Are Large Language Model Capabilities? A Case Study on BIG-bench Paper • 2305.14947 • Published May 24, 2023
Learning to Generate Task-Specific Adapters from Task Description Paper • 2101.00420 • Published Jan 2, 2021 • 2
Eliciting and Understanding Cross-Task Skills with Task-Level Mixture-of-Experts Paper • 2205.12701 • Published May 25, 2022 • 1