repo-context-repositioning

Enable language models to dynamically assign continuous position values to tokens based on contextual relevance instead of fixed linear positions. Implement learnable SwiGLU module extracting position information, integrate with RoPE for end-to-end optimization. Improves performance on noisy contexts and long-sequence tasks.

Independent GitHub snapshot · Automated classification · Original documentation governs