📰 Must-read papers and blogs on LLM based Long Context Modeling 🔥

[AAAI26] LongLLaDA: Unlocking Long Context Capabilities in Diffusion LLMs

Context-aware Biases for Length Extrapolation

[ICLR2025] ReAttention, a training-free approach to break the maximum context length in length extrapolation

Controlled experiments on recurrent-depth (looped) transformers — length extrapolation, adaptive test-time compute, and honest sub-1B negatives.

A constant-memory recurrent stream coupled to a growing .causal knowledge index, with a measured capacity threshold in the gated readout. Builds on GSSM.