Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

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

About

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.

Sean Welleck, Amanda Bertsch, Matthew Finlayson, Hailey Schoelkopf, Alex Xie, Graham Neubig, Ilia Kulikov, Zaid Harchaoui• 2024

Related benchmarks

TaskDatasetResultRank
Synthetic Grammar SynthesisSynthetic Grammar Synthesis a^n b^n c^n
Accuracy88.9
29
PlanningBlocks (Blocksworld)
Accuracy48.8
12
Combinatorial ReasoningGraph Coloring
Accuracy100
9
Combinatorial ReasoningSudoku 3x3
Accuracy96.7
9
Combinatorial ReasoningSudoku 4x4
Accuracy80
9
Synthetic Grammar SynthesisSynthetic Grammar Synthesis (a^m b^n c^m d^n)
Accuracy22.2
9
Showing 6 of 6 rows

Other info

Follow for update