Why Can't I Open My Drawer? Mitigating Object-Driven Shortcuts in Zero-Shot Compositional Action Recognition Paper • 2601.16211 • Published 26 days ago • 54
penfever/exp_rpt_curriculum-hard-qwen3.5-122b-131k-opencode-traces Viewer • Updated 21 days ago • 22 • 50 • 1
The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning Paper • 2606.29526 • Published about 1 month ago • 170
One Model, Many Latencies: Universal Speech Enhancement for Diverse Real-Time Applications Paper • 2606.25621 • Published Jun 24 • 22
SAE Interventions are Unreliable: Post-Intervention Recovery of Suppressed Behavior Paper • 2606.18322 • Published Jun 16 • 17
Imaginative Perception Tokens Enhance Spatial Reasoning in Multimodal Language Models Paper • 2606.03988 • Published Jun 3 • 126
One Click per Cell Type Suffices: Training-free Group Interaction for Cell Instance Segmentation Paper • 2605.29429 • Published May 28 • 8
Thinking Before Constraining: A Unified Decoding Framework for Large Language Models Paper • 2601.07525 • Published May 28 • 10
FlowLong: Inference-time Long Video Generation via Manifold-constrained Tweedie Matching Paper • 2605.20910 • Published May 20 • 29