Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning Paper • 2506.01939 • Published 13 days ago • 151
MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis Paper • 2502.18924 • Published Feb 26 • 13
Running 61 61 WebGPU Embedding Benchmark 🐠 Measure execution times of BERT models using WebGPU and WASM