11 citations · 41 across the 16 of their papers we have counts for
21 papers
Rethinking Pre-pretraining
Alexander Stern, Tim Rocktäschel, Acyr Locatelli +1
We report that allocating compute to synthetic pre-pretraining (PPT) never outperforms pure natural-language training in our compute-matched experiments. This holds across four PPT…
Why Speculative Decoding Works Better Than Predicted on Sparse MoE Models
Ekagra Ranjan, Komal K. Teru, Bharat Venkitesh +1
Speculative decoding (SD) accelerates autoregressive inference by leveraging the memory bound nature of decoding, verifying multiple draft tokens in a single forward pass. For mixt…
Legibility is Not Interpretability: Comparing Judged and Actual Importance in Chain-Of-Thought Reasoning
Kevin Du, Alexander Hoyle, Laura Ruis +1
Reasoning traces from chain-of-thought models appear to offer a legible window into how a model arrives at its answer. A growing body of work treats them as such, using LLM judges…
Efficient Benchmarking Is Just Feature Selection and Multiple Regression
Sam Bowyer, Acyr Locatelli, Kris Cao
Efficient benchmarking techniques aim to lower the computational cost of evaluating LLMs by predicting full benchmark scores using only a subset of a benchmark's questions. By refr…
Tiny Aya: Bridging Scale and Multilingual Depth
Alejandro R. Salamanca, Diana Abagyan, Daniel D'souza +23
Tiny Aya redefines what a small multilingual language model can achieve. Trained on 70 languages and refined through region-aware posttraining, it delivers state-of-the-art in tran…
One Tokenizer To Rule Them All: Emergent Language Plasticity via Multilingual Tokenizers
Diana Abagyan, Alejandro R. Salamanca, Andres Felipe Cruz-Salinas +6
Pretraining massively multilingual Large Language Models (LLMs) for many languages at once is challenging due to limited model capacity, scarce high-quality data, and compute const…
Aya Vision: Advancing the Frontier of Multilingual Multimodality
Saurabh Dash, Yiyang Nan, John Dang +22
Building multimodal language models is fundamentally challenging: it requires aligning vision and language modalities, curating high-quality instruction data, and avoiding the degr…
Command A: An Enterprise-Ready Large Language Model
Team Cohere, :, Aakanksha +227
In this report we describe the development of Command A, a powerful large language model purpose-built to excel at real-world enterprise use cases. Command A is an agent-optimised…
Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study
Menglong Cui, Pengzhi Gao, Wei Liu +2
Large language models (LLMs) have shown continuously improving multilingual capabilities, and even small-scale open-source models have demonstrated rapid performance enhancement. I…
Rope to Nope and Back Again: A New Hybrid Attention Strategy
Bowen Yang, Bharat Venkitesh, Dwarak Talupuru +4
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 (Ch…
Aya Expanse: Combining Research Breakthroughs for a New Multilingual Frontier
John Dang, Shivalika Singh, Daniel D'souza +42
We introduce the Aya Expanse model family, a new generation of 8B and 32B parameter multilingual language models, aiming to address the critical challenge of developing highly perf…
Procedural Knowledge in Pretraining Drives Reasoning in Large Language Models
Laura Ruis, Maximilian Mozes, Juhan Bae +7
The capabilities and limitations of Large Language Models have been sketched out in great detail in recent years, providing an intriguing yet conflicting picture. On the one hand,…
Understanding Likelihood Over-optimisation in Direct Alignment Algorithms
Zhengyan Shi, Sander Land, Acyr Locatelli +2
Direct Alignment Algorithms (DAAs), such as Direct Preference Optimisation (DPO) and Identity Preference Optimisation (IPO), have emerged as alternatives to online Reinforcement Le…
Nexus: Specialization meets Adaptability for Efficiently Training Mixture of Experts
Nikolas Gritsch, Qizhen Zhang, Acyr Locatelli +2
Efficiency, specialization, and adaptability to new data distributions are qualities that are hard to combine in current Large Language Models. The Mixture of Experts (MoE) archite…
To Code, or Not To Code? Exploring Impact of Code in Pre-training
Viraat Aryabumi, Yixuan Su, Raymond Ma +6
Including code in the pre-training data mixture, even for models not specifically designed for code, has become a common practice in LLMs pre-training. While there has been anecdot…
BAM! Just Like That: Simple and Efficient Parameter Upcycling for Mixture of Experts
Qizhen Zhang, Nikolas Gritsch, Dwaraknath Gnaneshwar +8
The Mixture of Experts (MoE) framework has become a popular architecture for large language models due to its superior performance over dense models. However, training MoEs from sc…
Aya 23: Open Weight Releases to Further Multilingual Progress
Viraat Aryabumi, John Dang, Dwarak Talupuru +18
This technical report introduces Aya 23, a family of multilingual language models. Aya 23 builds on the recent release of the Aya model (Üstün et al., 2024), focusing on pairing a…
SnapKV: LLM Knows What You are Looking for Before Generation
Yuhong Li, Yingbing Huang, Bowen Yang +6
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. Howe…
Pushing Mixture of Experts to the Limit: Extremely Parameter Efficient MoE for Instruction Tuning
Ted Zadouri, Ahmet Üstün, Arash Ahmadian +3
The Mixture of Experts (MoE) is a widely known neural architecture where an ensemble of specialized sub-models optimizes overall performance with a constant computational cost. How…
Exploring Low Rank Training of Deep Neural Networks
Siddhartha Rao Kamalakara, Acyr Locatelli, Bharat Venkitesh +3
Training deep neural networks in low rank, i.e. with factorised layers, is of particular interest to the community: it offers efficiency over unfactorised training in terms of both…
Regular cylindrical algebraic decomposition
J. H. Davenport, A. F. Locatelli, G. K. Sankaran
We show that a strong well-based cylindrical algebraic decomposition P of a bounded semi-algebraic set is a regular cell decomposition, in any dimension and independently of the me…