Pith. sign in

REVIEW 6 cited by

The Devil in Linear Transformer

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 2210.10340 v1 pith:KGBINX3H submitted 2022-10-19 cs.CL cs.LG

classification cs.CLcs.LG
keywords attentionlineargradientstransformertransformersdevildilutionexisting
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Linear transformers aim to reduce the quadratic space-time complexity of vanilla transformers. However, they usually suffer from degraded performances on various tasks and corpus. In this paper, we examine existing kernel-based linear transformers and identify two key issues that lead to such performance gaps: 1) unbounded gradients in the attention computation adversely impact the convergence of linear transformer models; 2) attention dilution which trivially distributes attention scores over long sequences while neglecting neighbouring structures. To address these issues, we first identify that the scaling of attention matrices is the devil in unbounded gradients, which turns out unnecessary in linear attention as we show theoretically and empirically. To this end, we propose a new linear attention that replaces the scaling operation with a normalization to stabilize gradients. For the issue of attention dilution, we leverage a diagonal attention to confine attention to only neighbouring tokens in early layers. Benefiting from the stable gradients and improved attention, our new linear transformer model, transNormer, demonstrates superior performance on text classification and language modeling tasks, as well as on the challenging Long-Range Arena benchmark, surpassing vanilla transformer and existing linear variants by a clear margin while being significantly more space-time efficient. The code is available at https://github.com/OpenNLPLab/Transnormer .

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

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

  1. DOSA: A Tree-Guided, Self-Regressive Framework for Long Document Structure Analysis

    cs.AI 2026-07 conditional novelty 6.0 of 10

    DOSA's tree-guided self-regressive chunking improves multi-page document hierarchy parsing by up to 4 F1 and 19 TEDS over prior methods on DocHieNet.

  2. MENTOR: A Metacognition-Driven Self-Evolution Framework for Uncovering and Mitigating Implicit Domain Risks in LLMs

    cs.AI 2025-11 conditional novelty 6.0 of 10

    A metacognition-driven self-evolution framework with activation steering reduces LLM jailbreak success rates on implicit domain risks from 57.8% to 4.6%.

  3. StegaFFD: Privacy-Preserving Face Forgery Detection via Fine-Grained Steganographic Domain Lifting

    cs.CV 2026-03 conditional novelty 5.0 of 10

    StegaFFD hides face images inside innocent cover photos and performs face-forgery detection directly on the stego image, preserving accuracy while hiding the face.

  4. Elucidating the Design Space of Decay in Linear Attention

    cs.CL 2025-09 conditional novelty 5.0 of 10

    A controlled study of decay in linear attention finds median decay near 0.8 works best, vector decay generally beats scalar decay, and RoPE/TPE give little benefit for models with sub-unity decay.

  5. Transolver++: An Accurate Neural Solver for PDEs on Million-Scale Geometries

    cs.LG 2025-02 conditional novelty 5.0 of 10

    Transolver++ scales a physics-state attention model to 2.5 million mesh points via local adaptive temperatures, Gumbel-Softmax slice assignment, and AllReduce communication of physics states.

  6. Position: The Future of Bayesian Prediction Is Prior-Fitted

    cs.LG 2025-05 conditional novelty 4.0 of 10

    PFNs, which amortize Bayesian inference by training on datasets sampled from a prior, are likely to supersede MCMC and variational inference for most prediction tasks, the authors argue.

Pith tools