REVIEW 8 cited by
Fourier Position Embedding: Enhancing Attention's Periodic Extension for Length Generalization
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Fourier Position Embedding: Enhancing Attention's Periodic Extension for Length Generalization
read the original abstract
Extending the context length of Language Models (LMs) by improving Rotary Position Embedding (RoPE) has become a trend. While prior works mainly address RoPE's limitations within attention, this paper uncovers the adverse effects on length generalization from nearly all parts of LMs. Using Discrete Signal Processing theory, we show that RoPE enables periodic attention by implicitly achieving Non-Uniform Discrete Fourier Transform. However, this periodicity is undermined by the spectrum damage caused by: 1) linear layers and activation functions; 2) insufficiently trained frequency components brought by time-domain truncation. Building on our observations, we propose Fourier Position Embedding (FoPE), which enhances attention's frequency-domain properties to improve both its periodic extension and length generalization. FoPE constructs \textit{Fourier Series} and zero-outs the destructive frequency components, increasing model robustness against the spectrum damage. Experiments across various model scales and benchmarks show that, within varying context windows, FoPE maintains a more stable performance compared to other baselines. Several analyses and ablations bring further support to our method and theoretical modeling.
Forward citations
Cited by 8 Pith papers
-
LazyAttention: Efficient Retrieval-Augmented Generation with Deferred Positional Encoding
LazyAttention kernelizes deferred positional encoding to enable zero-copy, position-agnostic KV cache reuse, delivering 1.37× lower TTFT and 1.40× higher throughput than Block-Attention under skewed document distribut...
-
ClockRoPE: Random Fourier Rotations for Temporal Routine Modeling
Random Fourier Rotations let transformer position encodings approximate any positive-definite attention kernel; ClockRoPE applies this to model daily/weekly routines in sequential recommendation.
-
Weierstrass Positional Encoding for Vision Transformers
WePE encodes 2D patch positions in Vision Transformers via Weierstrass elliptic functions on the complex plane to exploit double periodicity and derive relative positions algebraically.
-
Convergent Evolution: How Different Language Models Learn Similar Number Representations
Diverse language models converge on similar periodic number features with a two-tier hierarchy of Fourier sparsity and geometric separability, acquired via language co-occurrences or multi-token arithmetic.
-
UCAN: Unified Convolutional Attention Network for Expansive Receptive Fields in Lightweight Super-Resolution
UCAN unifies window-based spatial attention and Hedgehog Attention with a distillation-based large-kernel module and cross-layer sharing to deliver competitive PSNR at low MACs in lightweight super-resolution.
-
Kimi Linear: An Expressive, Efficient Attention Architecture
Kimi Linear hybridizes linear attention with a new KDA module to beat full attention on tasks while slashing KV cache by 75% and speeding decoding up to 6x.
-
Mitigating Coordinate Prediction Bias from Positional Encoding Failures
VPSG corrects predictable directional coordinate biases in MLLMs by shuffling visual positional encodings to isolate unconditioned tendencies and steering digit decoding with a lightweight finite-state machine, yieldi...
-
Mix-MoE: Improving Multilingual Machine Translation of Large Language Models through Mixed MoEs
Mix-MoE applies separate LM and MT expert groups in two post-pretraining stages with Fourier-enhanced routing to reduce parameter interference and improve multilingual MT over baselines.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.