Scale AI · 10 records

dateidtitleauthorsvenue↑
2026-10-202610.17724Evidence-First Agentic VerificationReza Sanayei et al.arXiv21
2026-10-122610.08966Humanity's Sixth Sense: Benchmarking Intuitive Visual Reasoning in Multimodal ModelsXingang Guo et al.arXiv12
2026-09-212609.12623SteerDuplex: Steerable Duplex Speech Dialogue ModelsUtkarsh Tyagi et al.arXiv5
2026-09-152609.13463Root-Cause Attribution Is a Search Problem: Continual Search for Long-Horizon Agent FailuresHarsh Raj et al.arXiv5
2026-09-142609.10824Studying Without a Syllabus: Task-Agnostic Environment PreprocessingVinay Samuel et al.arXiv5
2026-08-072608.06301HarnessOpt-Bench: Evaluating LLMs at Harness OptimizationVarun Ursekar et al.arXiv15
2026-08-042607.28802Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent FailuresHarsh Raj et al.arXiv9
2026-07-012606.30573SWE-INTERACT: Reimagining SWE Benchmarks as User-Driven Long-Horizon Coding SessionsMohit Raghavendra et al.arXiv7
2026-05-202605.20164Not Every Rubric Teaches Equally: Policy-Aware Rubric Rewards for RLVRUtkarsh Tyagi et al.arXiv5
2026-05-052604.09408HiL-Bench (Human-in-Loop Benchmark): Do Agents Know When to Ask for Help?Mohamed Elfeki et al.arXiv4