AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

FoNE: Precise Single-Token Number Embeddings via Fourier Features

Tianyi Zhou, Deqing Fu, Mahdi Soltanolkotabi, Robin Jia, Vatsal Sharan

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Pruning and circuits in LLMs · 13 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use single-token FoNE instead of digit-wise tokenstransformer · 3-digit multiplicationaccuracy (%): 81.21 -> 98.56Improved
use single-token FoNE instead of subword tokenstransformer · 3-digit multiplicationaccuracy (%): 8.05 -> 98.56Improved
use FoNE instead of xValtransformer · 6-digit decimal additionaccuracy (%): 0.44 -> 100Improved
use FoNE instead of digit-wise tokenstransformer · 6-digit decimal additiontraining time: 11'48'' -> 3'18''Improved
use FoNE (one token) instead of digit-wisetransformer · 6-digit decimal additiontokens per number: 7 -> 1Improved
replace linear layer with zero padding to match embedding dimension (ablation)FoNE · Multiplicationaccuracy (%): 99.95 -> 99.91No clear effect
add periods 2 and 5 to period 10 (ablation)FoNE · Multiplicationaccuracy (%): 99.95 -> 99.99No clear effect
use period 5 instead of 10 (ablation)FoNE · Decimal additionaccuracy (%): 100 -> 1.52Hurt
use period 7 instead of 10 (ablation)FoNE · Decimal additionaccuracy (%): 100 -> 3.64Hurt
add simplified FoNE to Llama-3.1-1B continual pretrainingLlama-3.1-1B · 4-digit subtractionaccuracy (%): 29.60 -> 39.90Improved
add simplified FoNE to Llama-3.1-1BLlama-3.1-1B · MMLUaccuracy (%): 38.10 -> 38.21No clear effect
use FoNE tokenization in GPT-2 117M pretrainingGPT-2 117M · FineWeb text-token perplexityperplexity: 67.390 -> 46.856Improved

Later papers that compared with this method

1 later accepted paper reports FoNE as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy