Your LLM Knows the Future: Uncovering Its Multi-Token Prediction Potential

Autoregressive language models are constrained by their inherently sequential nature, generating one token at a time. This paradigm limits inference speed and parallelism, especially during later stages of generation when the direction and semantics of text are relatively certain. In this work, we propose a novel framework that leverages the inherent knowledge of vanilla autoregressive language models about future tokens, combining techniques to realize this potential and enable simultaneous prediction of multiple subsequent tokens. Our approach introduces several key innovations: (1) a masked-input formulation where multiple future tokens are jointly predicted from a common prefix; (2) a gated LoRA formulation that preserves the original LLM's functionality, while equipping it for multi-token prediction; (3) a lightweight, learnable sampler module that generates coherent sequences from the predicted future tokens; (4) a set of auxiliary training losses, including a consistency loss, to enhance the coherence and accuracy of jointly generated tokens; and (5) a speculative generation strategy that expands tokens quadratically in the future while maintaining high fidelity. Our method achieves significant speedups through supervised fine-tuning on pretrained models. For example, it generates code and math nearly 5x faster, and improves general chat and knowledge tasks by almost 2.5x. These gains come without any loss in quality.

transformers/src/transformers/models/auto/tokenization_auto.py at 1d5296f2de31ecee4fdd789dc2b2fe068b320284 · huggingface/transformers

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training. - huggingface/transformers

Tokenization: A Survey for Modern NLP

While modern language models take raw text as their input and produce raw text as output, they do not operate over text directly. Hidden in the very first step of language model pipelines is...

Language models struggle with compartmentalization

In the training data used by large language models (LLMs), the same latent concept is often presented in multiple distinct ways: the same facts appear in English and Swahili; many functions can be expressed in both Python and Haskell; we can express propositions in both formal and natural language. We show that LLMs can exhibit compartmentalization, where they fail to identify and share statistical strength between distinct presentations of unified concepts. In the worst case, LLMs simply learn parallel internal representations of each presentation of the concept, saturating model capacity with redundancies and decreasing sample efficiency with the number of such presentations. We also demonstrate that synthetic parallel data can fail to improve this despite being easily learned itself. Under this framework, we find that, for small models, early multilingual learning is nearly entirely compartmentalized. Finally, all interventions that we study exhibit a phase transition in which their effectiveness depends on the number of distinct presentations, suggesting that the language modeling objective may only inconsistently unify representations.

What are embeddings?

A deep-dive into machine learning embeddings.

Machine Learning  |  Google for Developers

How linguistics learned to stop worrying and love the language models

Language models (LMs) can produce fluent, grammatical text. Nonetheless, some maintain that language models don’t really learn language and also, even if they did, that would not be informative for the study of human learning and processing. On the other side, there have been claims that the success of LMs obviates the need for studying linguistic theory and structure. We argue that both extremes are wrong. LMs can contribute to fundamental questions about linguistic structure, language processing, and learning. They force us to rethink arguments and ways of thinking that have been foundational in linguistics. While they do not replace linguistic structure and theory, they serve as model systems and working proofs of concept for gradient, usage-based approaches to language. We offer an optimistic take on the relationship between language models and linguistics.

REKI: Registering Knowledge in VLAs

REKI, a 2.4B vision-language-action model pretrained on public data only, with a memory of its recent past at negligible cost.

Stanford CS336 | Language Modeling from Scratch (Spring 2025 Archive)

Archived course website for Stanford CS336: Language Modeling from Scratch (Spring 2025), including schedule, assignments, logistics, and materials.

Introducing beginners to the mechanics of machine learning – Miriam Posner

Every year, I spend some time introducing students to the mechanics of machine learning with neural nets. I definitely don’t go into great depth; I usually only have one class for this. But I try to unpack at least some of the major concepts, so that ML isn’t quite such a black box.

Recursive Language Models: the paradigm of 2026

How we plan to manage extremely long contexts

Generalization Dynamics of LM Pre-training

Generalization Dynamics of LM Pre-Training shows that language models can switch abruptly between shallow pattern-following and more transferable computation during pre-training, even when loss and...