…gradual architecture transitions lead to better optimization trajectories. Putting it all together, we compress Qwen3-Next-80A3B to a 23A2B model that retains competitive performance. These results offer practical guidance for efficient…
…To put it directly, RL fundamentally optimizes the recall of latent knowledge. 2️⃣ The unexpected contribution of 0/128 samples: Remarkably, ~83% of the performance jump is driven by training on the…
…Reaches competitive performance in just 80 epochs compared to 800+ epochs for previous methods. Diffusion-Friendly Manifold: Explicitly optimizes three key geometric properties: Spatial Structure Coherence, Local Manifold Continuity, and Global Semantic…
…Enhancing Group Relative Policy Optimization with Verifiable Process Supervision for Effective Reasoning (2026) Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes (2026) Self-Induced Outcome Potential: Turn-Level Credit Assignment…
…Yihong Liu , , , Abstract COPSD transfers high-resource language model reasoning behavior to low-resource languages using self-distillation with crosslingual context, improving mathematical reasoning performance. AI-generated summary Large language models (LLMs…
…These developments lead to strong performance in multimodal coding, visual tool use , and framework-based agentic tasks, while preserving competitive text-only coding capability. More importantly, our development process offers practical insights…
…We present RewardHarness, a self-evolving agentic reward framework that reframes reward modeling as context evolution rather than weight optimization. Instead of learning from large-scale annotations, RewardHarness aligns with human preferences…
…Yu-Hsi Chen , Abstract CPCANet is a domain generalization framework that uses Common Principal Component Analysis to discover structured domain-invariant subspaces through differentiable neural layers, achieving state-of-the-art performance…
…overhead (1.2% in the optimized release). Empirical results show that ESamp significantly boosts the Pass@k efficiency of reasoning models, showing superior or comparable performance to strong stochastic and heuristic baselines…
…Thinking with 4D Imagery for Dynamic Spatial Understanding Published on May 7 Submitted by jankin on May 11 Tsinghua University Authors: , , , Xiang An , , , , , , , , Abstract 4DThinker enables vision-language models to perform dynamic…