From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Models

TL;DR

This paper surveys inference-time algorithms for large language models, focusing on token-level, meta-generation, and efficiency methods, highlighting their roles in improving generation quality and speed during inference.

Contribution

It unifies diverse inference algorithms under a common formalism, bridging NLP, LLMs, and ML systems to advance understanding of inference-time improvements.

Findings

Token-level algorithms improve sampling quality.

Meta-generation incorporates external knowledge.

Efficiency methods reduce token costs.

Abstract

One of the most striking findings in modern research on large language models (LLMs) is that scaling up compute during training leads to better results. However, less attention has been given to the benefits of scaling compute during inference. This survey focuses on these inference-time approaches. We explore three areas under a unified mathematical formalism: token-level generation algorithms, meta-generation algorithms, and efficient generation. Token-level generation algorithms, often called decoding algorithms, operate by sampling a single token at a time or constructing a token-level search space and then selecting an output. These methods typically assume access to a language model's logits, next-token distributions, or probability scores. Meta-generation algorithms work on partial or full sequences, incorporating domain knowledge, enabling backtracking, and integrating external information. Efficient generation methods aim to reduce token costs and improve the speed of generation. Our survey unifies perspectives from three research communities: traditional natural language processing, modern LLMs, and machine learning systems.