Related papers: Kinetics: Rethinking Test-Time Scaling Laws
Test-time scaling (TTS) improves large language models (LLMs) by allocating additional compute at inference time. In practice, TTS is often achieved through parallel scaling: generating multiple candidate responses and selecting the best…
In recent years, the state-of-the-art in deep learning has been dominated by very large models that have been pre-trained on vast amounts of data. The paradigm is very simple: investing more computational resources (optimally) leads to…
The self-attention mechanism traditionally relies on the softmax operator, necessitating positional embeddings like RoPE, or position biases to account for token order. But current methods using still face length generalisation challenges.…
As large language models (LLMs) and visual language models (VLMs) grow in scale and application, attention mechanisms have become a central computational bottleneck due to their high memory and time complexity. While many efficient…
Large language models (LLMs) demonstrate strong capabilities across a wide range of complex tasks and are increasingly deployed at scale, placing significant demands on inference efficiency. Prior work typically decomposes inference into…
Recently, Test-Time Scaling (TTS) has gained increasing attention for improving LLM reasoning performance at test time without retraining the model. A notable TTS technique is Self-Consistency (SC), which generates multiple reasoning chains…
Layer-wise quantization is a key technique for efficiently compressing large models without expensive retraining. Previous methods typically quantize the weights of each layer by "uniformly" optimizing the layer reconstruction loss across…
Scale has been a major driving force in improving machine learning performance, and understanding scaling laws is essential for strategic planning for a sustainable model quality performance growth, long-term resource planning and…
While scaling laws for large language models (LLMs) during pre-training have been extensively studied, their behavior under reinforcement learning (RL) post-training remains largely unexplored. This paper presents a systematic empirical…
Leveraging attention sparsity to accelerate long-context large language models (LLMs) has been a hot research topic. However, current algorithms such as sparse attention or key-value (KV) cache compression tend to use a fixed budget, which…
Large language models (LLMs) now support extremely long context windows, but the quadratic complexity of vanilla attention results in significantly long Time-to-First-Token (TTFT) latency. Existing approaches to address this complexity…
Code Large Language Models (LLMs) are revolutionizing software engineering. However, scaling laws that guide the efficient training are predominantly analyzed on Natural Language (NL). Given the fundamental differences like strict syntax…
Transformer-based language models have found many diverse applications requiring them to process sequences of increasing length. For these applications, the causal self-attention -- which is the only component scaling quadratically w.r.t.…
Adaptive sample size re-estimation (SSR) is a well-established strategy for improving the efficiency and flexibility of clinical trials. Its central challenge is determining whether, and by how much, to increase the sample size at an…
Pre-trained models have become indispensable for efficiently building models across a broad spectrum of downstream tasks. The advantages of pre-trained models have been highlighted by empirical studies on scaling laws, which demonstrate…
An efficient attention implementation is essential for large models due to its quadratic time complexity. Fortunately, attention commonly exhibits sparsity, i.e., many values in the attention map are near zero, allowing for the omission of…
Current scaling laws for visual AI models focus predominantly on large-scale pretraining, leaving a critical gap in understanding how performance scales for data-constrained downstream tasks. To address this limitation, this paper…
Cross-encoders are effective passage and document re-rankers but less efficient than other neural or classic retrieval models. A few previous studies have applied windowed self-attention to make cross-encoders more efficient. However, these…
Recently, scaling test-time compute on Large Language Models (LLM) has garnered wide attention. However, there has been limited investigation of how various reasoning prompting strategies perform as scaling. In this paper, we focus on a…
Scaling the context length of large language models (LLMs) offers significant benefits but is computationally expensive. This expense stems primarily from the self-attention mechanism, whose $O(N^2)$ complexity with respect to sequence…