Keyword Harnesses Fail Open: A Cheap Diagnostic Ladder for Tool-Use Claims in Small Language Models Paper • 2610.02142 • Published 2 days ago • 2
Running Featured 400 Jev Decision Index 🔬 400 Benchmarks and news on various repros of TypeSafe's Jev
ActiveSaddler: Automated Curriculum Learning for Agent Harness Optimization Paper • 2610.00906 • Published 2 days ago • 50
Keyword Harnesses Fail Open: A Cheap Diagnostic Ladder for Tool-Use Claims in Small Language Models Paper • 2610.02142 • Published 2 days ago • 2
Pitwall: Faithful Natural-Language Race-Strategy Briefings from a Calibrated Real-Time Monte Carlo Engine Paper • 2607.06495 • Published Jul 7
VectraYX-Vision-1B: A Sub-2B Spanish/LATAM Cybersecurity Vision-Language Model with Structured Visual Reasoning and Native Tool Use Paper • 2608.08477 • Published Aug 9 • 6