Pith. sign in

REVIEW 17 cited by

Round and Round We Go! What makes Rotary Positional Encodings useful?

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

arxiv 2410.06205 v3 pith:5MGWGTP3 submitted 2024-10-08 cs.CL cs.LG

Round and Round We Go! What makes Rotary Positional Encodings useful?

classification cs.CL cs.LG
keywords ropellmspositionalencodingsgemmausedattentionbelieve
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Positional Encodings (PEs) are a critical component of Transformer-based Large Language Models (LLMs), providing the attention mechanism with important sequence-position information. One of the most popular types of encoding used today in LLMs are Rotary Positional Encodings (RoPE), that rotate the queries and keys based on their relative distance. A common belief is that RoPE is useful because it helps to decay token dependency as relative distance increases. In this work, we argue that this is unlikely to be the core reason. We study the internals of a trained Gemma 7B model to understand how RoPE is being used at a mechanical level. We find that Gemma learns to use RoPE to construct robust "positional" attention patterns by exploiting the highest frequencies. We also find that, in general, Gemma greatly prefers to use the lowest frequencies of RoPE, which we suspect are used to carry semantic information. We mathematically prove interesting behaviours of RoPE and conduct experiments to verify our findings, proposing a modification of RoPE that fixes some highlighted issues and improves performance. We believe that this work represents an interesting step in better understanding PEs in LLMs, which we believe holds crucial value for scaling LLMs to large sizes and context lengths.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 17 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Fingerprint, Not Blueprint: How Positional Schemes Set the Default Spectral Algebra of Attention

    cs.LG 2026-07 conditional novelty 7.0

    Positional schemes set the default spectral algebra of attention heads: previous-token heads are rotational under RoPE and content-like under absolute/ALiBi, as a post-function fingerprint rather than a hard constraint.

  2. Why Muon Outperforms Adam: A Curvature Perspective

    cs.LG 2026-06 conditional novelty 7.0

    Muon outperforms Adam by reducing curvature penalty via lower Normalized Directional Sharpness, as shown via Taylor approximation on LLM training and proven on stylized quadratic problems with heterogeneous curvature.

  3. Rotation-Aligned Key Channel Pruning for Efficient Vision-Language Model Inference

    cs.CV 2026-05 unverdicted novelty 7.0

    RotateK uses online PCA-based rotation to align token-dependent key channel importance into a shared subspace, enabling accurate head-wise structured pruning and faster decoding in VLMs compared to prior token or chan...

  4. RoPE Distinguishes Neither Positions Nor Tokens in Long Contexts, Provably

    cs.CL 2026-05 conditional novelty 7.0

    Proves that RoPE attention loses locality bias and token distinction in long contexts, approaching random behavior independent of content.

  5. Dual Triangle Attention: Effective Bidirectional Attention Without Positional Embeddings

    q-bio.QM 2026-04 unverdicted novelty 7.0

    Dual Triangle Attention achieves effective bidirectional attention with built-in positional inductive bias via dual triangular masks, outperforming standard bidirectional attention on position-sensitive tasks and show...

  6. Group Representational Position Encoding

    cs.LG 2025-12 unverdicted novelty 7.0

    GRAPE unifies RoPE and ALiBi as special cases of group actions on positions, providing a principled design space for positional encodings via SO(d) rotations and GL unipotent transformations.

  7. SHERPA: Seam-aware Harmonized ERP Adaptation for Open-Domain 360$^\circ$ Panorama Generation

    cs.CV 2026-06 unverdicted novelty 6.0

    SHERPA adapts diffusion models for open-domain 360° ERP panorama generation via frequency-selective Circular RoPE, circular latent encoding, FFN adapters, and dual-path training with yaw consistency.

  8. RoVE: Rotary Value Embeddings Attention for Relative Position-dependent Value Pathways

    cs.LG 2026-06 unverdicted novelty 6.0

    RoVE rotates value embeddings simultaneously with keys in attention to make values position-dependent, reframing RoPE as attentive convolution and reporting gains on long-context tasks in 124M and 354M GPT-2 models.

  9. Stateful Reasoning via Insight Replay

    cs.AI 2026-05 conditional novelty 6.0

    InsightReplay improves LLM accuracy on reasoning benchmarks by extracting and replaying critical insights to maintain their accessibility during extended chain-of-thought generation.

  10. Stateful Reasoning via Insight Replay

    cs.AI 2026-05 unverdicted novelty 6.0

    InsightReplay improves long CoT reasoning by extracting critical insights from the trace and replaying them near the active frontier, delivering +1.65 average accuracy gain across 24 model-benchmark settings.

  11. When Attention Closes: How LLMs Lose the Thread in Multi-Turn Interaction

    cs.AI 2026-05 unverdicted novelty 6.0

    Attention to goal tokens declines in multi-turn LLM interactions while residual representations often retain decodable goal information, and the gap between these predicts whether goal-conditioned behavior survives.

  12. Remember to Forget: Gated Adaptive Positional Encoding

    cs.LG 2026-05 unverdicted novelty 6.0

    GAPE augments RoPE with query- and key-dependent gates to stabilize attention and improve long-context performance in language models.

  13. A3 : an Analytical Low-Rank Approximation Framework for Attention

    cs.CL 2025-05 conditional novelty 6.0

    A3 splits Transformer layers into QK, OV, and MLP components and derives analytical low-rank approximations that reduce hidden dimensions while minimizing each component's functional loss, yielding better perplexity t...

  14. TabICL: A Tabular Foundation Model for In-Context Learning on Large Data

    cs.LG 2025-02 unverdicted novelty 6.0

    TabICL scales in-context learning to large tabular data via column-then-row attention for row embeddings followed by a transformer, matching TabPFNv2 speed and performance while outperforming it and CatBoost on datase...

  15. LTX-Video: Realtime Video Latent Diffusion

    cs.CV 2024-12 conditional novelty 6.0

    LTX-Video integrates Video-VAE and transformer for 1:192 latent compression and real-time video diffusion by moving patchifying to the VAE and letting the decoder finish denoising in pixel space.

  16. Adaptive 3D-RoPE: Physics-Aligned Rotary Positional Encoding for Wireless Foundation Models

    eess.SP 2026-05 unverdicted novelty 5.0

    Adaptive 3D-RoPE adapts rotary positional encoding to wireless channel physics via learnable 3D frequencies and dynamic CSI control, yielding up to 10.7 dB NMSE gains in scale extrapolation and 1 dB in zero-shot tasks.

  17. LongAct: Harnessing Intrinsic Activation Patterns for Long-Context Reinforcement Learning

    cs.LG 2026-04 unverdicted novelty 5.0

    LongAct uses saliency from high-magnitude activations to guide sparse weight updates in long-context RL, yielding about 8% gains on LongBench v2 across multiple algorithms.