LLM Digest

Story

arxiv_cs_cl · Nov 4, 2026 · paper

Source brief

LooM: Loop-Invariant Memory for Causal Encoder-Decoder Language Models

arxiv.orgNov 4, 2026
original source linked

In brief

Looped Transformers increase effective depth by re-executing shared layers, but three limitations hinder their adoption: memory grows with the loop count, compute can only be adjusted in whole-block steps, and the loop count is hard to change after training. We propose LooM (Loop

Read the original at arxiv.org →Open in live feed Read that day’s brief

LooM: Loop-Invariant Memory for Causal Encoder-Decoder Language Models

Earlier in this thread 4 items

TraceRouter++: Causal Semantic Routing Realignment via Optimal Transport

CausalXRL: Explainable Reinforcement Learning through Causal Graph Reasoning

Causal Attention on CPU: A Cache-Aware and Load-Balanced Scheduling Kernel for Prefill

Causal Chain of Thought: Mitigating Attention Leakage via Context Isolation