StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling
Paper • 2608.15089 • Published • 445
sha7dow
sha7doww
AI & ML interests
None yet
Recent Activity
upvoted a paper 19 days ago
Modular TTT: Rethinking Test-Time Training as Composable Modules upvoted a paper 19 days ago
Characterizing the Quality Profile of AI-Generated C++ in ProductionOrganizations
None yet
upvoted a paper 5 days ago
upvoted 3
papers 19 days ago
Modular TTT: Rethinking Test-Time Training as Composable Modules
Paper • 2608.07110 • Published • 9
Characterizing the Quality Profile of AI-Generated C++ in Production
Paper • 2608.06640 • Published • 11
The Optimizer Is the Agent: Reasoning-Driven Search across Prompts, Programs, and ML Workflows
Paper • 2608.06714 • Published • 10