Long-context large language models (LLMs) have achieved remarkable advancements, driven by techniques like Rotary Position Embedding (RoPE) (Su et al., 2023) and its extensions (Chen et al., 2023; Liu et al., 2024c; Peng et al., 2023). By adjusting RoPE parameters and incorporating training data with extended contexts, we can train performant models with considerably longer input sequences. However, existing RoPE-based methods exhibit performance limitations when applied to extended context lengths. This paper presents a comprehensive analysis of various attention mechanisms, including RoPE, No Positional Embedding (NoPE), and Query-Key Normalization (QK-Norm), identifying their strengths and shortcomings in long-context modeling. Our investigation identifies distinctive attention patterns in these methods and highlights their impact on long-context performance, providing valuable insights for architectural design. on long context performance, providing valuable insights for architectural design. Building on these findings, we propose a novel architecture featuring a hybrid attention mechanism that integrates global and local attention spans. This design not only surpasses conventional RoPE-based transformer models with full attention in both long and short context tasks but also delivers substantial efficiency gains during training and inference.

Large Language Models (LLMs) have made remarkable progress in processing extensive contexts, with the Key-Value (KV) cache playing a vital role in enhancing their performance. However, the growth of the KV cache in response to increasing input length poses challenges to memory and time efficiency. To address this problem, this paper introduces SnapKV, an innovative and fine-tuning-free approach that efficiently minimizes KV cache size while still delivering comparable performance in real-world applications. We discover that each attention head in the model consistently focuses on specific prompt attention features during generation. Meanwhile, this robust pattern can be obtained from an `observation' window located at the end of the prompts. Drawing on this insight, SnapKV automatically compresses KV caches by selecting clustered important KV positions for each attention head. Our approach significantly reduces the growing computational overhead and memory footprint when processing long input sequences. Specifically, SnapKV achieves a consistent decoding speed with a 3.6x increase in generation speed and an 8.2x enhancement in memory efficiency compared to baseline when processing inputs of 16K tokens. At the same time, it maintains comparable performance to baseline models across 16 long sequence datasets. Moreover, SnapKV can process up to 380K context tokens on a single A100-80GB GPU using HuggingFace implementation with minor changes, exhibiting only a negligible accuracy drop in the Needle-in-a-Haystack test. Further comprehensive studies suggest SnapKV's potential for practical applications.

Mixture of Experts (MoE) framework has become a popular architecture for large language models due to its superior performance compared to dense models. However, training MoEs from scratch in a large-scale regime is prohibitively expensive. Previous work addresses this challenge by independently training multiple dense expert models and using them to initialize an MoE. In particular, state-of-the-art approaches initialize MoE layers using experts' feed-forward parameters while merging all other parameters, limiting the advantages of the specialized dense models when upcycling them as MoEs. We propose BAM (Branch-Attend-Mix), a simple yet effective improvement to MoE training. BAM makes full use of specialized dense models by not only using their feed-forward network (FFN) to initialize the MoE layers but also leveraging experts' attention weights fully by leveraging them as mixture-of-attention (MoA) layers. We explore two methods for upcycling MoA layers: 1) initializing separate attention experts from dense models including key, value, and query matrices; and 2) initializing only Q projections while sharing key-value pairs across all experts to facilitate efficient inference. Our experiments using seed models ranging from 590 million to 2 billion parameters show that our approach outperforms state-of-the-art approaches under the same data and compute budget in both perplexity and downstream tasks evaluations, confirming the effectiveness of BAM.

Emergent properties have been widely adopted as a term to describe behavior not present in smaller models but observed in larger models (Wei et al., 2022a). Recent work suggests that the trade-off incurred by quantization is also an emergent property, with sharp drops in performance in models over 6B parameters. In this work, we ask _are quantization cliffs in performance solely a factor of scale?_ Against a backdrop of increased research focus on why certain emergent properties surface at scale, this work provides a useful counter-example. We posit that it is possible to optimize for a quantization friendly training recipe that suppresses large activation magnitude outliers. Here, we find that outlier dimensions are not an inherent product of scale, but rather sensitive to the optimization conditions present during pre-training. This both opens up directions for more efficient quantization, and poses the question of whether other emergent properties are inherent or can be altered and conditioned by optimization and architecture design choices. We successfully quantize models ranging in size from 410M to 52B with minimal degradation in performance.

We propose smart roaming (SR), a cooperation technique between cellular telcos operating in the same region, that enables users to roam for performance reasons (even if they are covered by their operator). Within a region, base stations of different operators are sometimes co-located but, in that case, the sectors are rarely aligned. SR leverages spatial diversity to enhance spectrum usage efficiency. Simply put, an edge user of an operator might be a “good user” for another one. This paper answers the following research questions: 1) Can significant gain be obtained with SR? 2) What are the factors that affect the gain? 3) How to manage operator heterogeneity to avoid that a large operator cross subsidizes a smaller one? and 4) How to implement and manage SR in an online fashion while keeping the signaling information manageable? We answer the first three questions by proposing a snapshot model for the downlink that shows that SR can indeed provide significant gain without yielding cross-subsidies if done properly. We then propose two schemes to implement and manage SR online and evaluate them via extensive simulations.