How an LLM edited 0.016% of Qwen3.5-0.8B’s weights to break its reasoning loops: directional ablation from the refusal-direction work, then pushed further with a strength α and 64 selected neurons per layer.
Softmax Collapse and Softmax Cancellation: two floating-point pitfalls in Softmax and Cross-Entropy, and mathematically equivalent fixes.
Rotation matrices, complex numbers and RoPE, and the 6D and 9D rotation representations used in neural networks.
Why does training loss creep up within each epoch and drop at the next? Not the optimizer, but what the loss measures: the expected gap since a sample was last trained on follows a sawtooth, and it explains every sampling ablation.
Clean, minimal PyTorch implementations of diffusion and flow models, from 1000-step DDPM to 1-step generators, on one backbone with one set of conventions.