view article Article Wringer: Fill, then Wring — a 4B Reasoning Model at 2.655 Bits per Weight with 93.7% of Its Benchmark Scores wcamon • 6 days ago
circus-0.3: Elicit, Attribute, Repair Collection Paper and minimized evidence for circus-0.3 agentic binding diagnosis and GSPO repair. Models remain gated. • 2 items • Updated 14 days ago