{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/sparse-attention-with-linear-units","title":"Sparse Attention with Linear Units","arxiv_id":"2104.07012","date":"2021-04-14","proceeding":"EMNLP 2021 11","authors":["Biao Zhang","Ivan Titov","Rico Sennrich"],"abstract":"Recently, it has been argued that encoder-decoder models can be made more interpretable by replacing the softmax function in the attention with its sparse variants. In this work, we introduce a novel, simple method for achieving sparsity in attention: we replace the softmax activation with a ReLU, and show that sparsity naturally emerges from such a formulation. Training stability is achieved with layer normalization with either a specialized initialization or an additional gating function. Our model, which we call Rectified Linear Attention (ReLA), is easy to implement and more efficient than previously proposed sparse attention mechanisms. We apply ReLA to the Transformer and conduct experiments on five machine translation tasks. ReLA achieves translation performance comparable to several strong baselines, with training and decoding speed similar to that of the vanilla attention. Our analysis shows that ReLA delivers high sparsity rate and head diversity, and the induced cross attention achieves better accuracy with respect to source-target word alignment than recent sparsified softmax-based models. Intriguingly, ReLA heads also learn to attend to nothing (i.e. 'switch off') for some queries, which is not possible with sparsified softmax alternatives.","url_abs":"https://arxiv.org/abs/2104.07012v2","url_pdf":"https://arxiv.org/pdf/2104.07012v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"sparse-attention-with-linear-units","repo_url":"https://github.com/bzhangGo/zero","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"sparse-attention-with-linear-units","repo_url":"https://github.com/lucidrains/FLASH-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"sparse-attention-with-linear-units","repo_url":"https://github.com/rishikksh20/rectified-linear-attention","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"diversity","task_name":"Diversity"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"translation","task_name":"Translation"},{"task_slug":"word-alignment","task_name":"Word Alignment"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2104.07012","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2104.07012"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lucidrains/FLASH-pytorch","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/bzhangGo/zero","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/rishikksh20/rectified-linear-attention","reach":null}],"summary":{"ran":3,"unverified":2},"by_repo_kind":{"official":{"samples":2,"ran":0,"repositories":1},"listed":{"samples":3,"ran":3,"repositories":2}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"cc210ef49adb120a","entry":"RMSNorm","repo":"rishikksh20/rectified-linear-attention","repo_kind":"listed","path":"attention.py","file_url":"https://github.com/rishikksh20/rectified-linear-attention/blob/HEAD/attention.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"cc210ef49adb120a"}},{"code_sha256_prefix":"78909de648a3fba3","entry":"ReLUSquared","repo":"lucidrains/FLASH-pytorch","repo_kind":"listed","path":"flash_pytorch/flash_pytorch.py","file_url":"https://github.com/lucidrains/FLASH-pytorch/blob/HEAD/flash_pytorch/flash_pytorch.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"78909de648a3fba3"}},{"code_sha256_prefix":"ce8755a66fa65637","entry":"RectifiedLinearAttention","repo":"rishikksh20/rectified-linear-attention","repo_kind":"listed","path":"attention.py","file_url":"https://github.com/rishikksh20/rectified-linear-attention/blob/HEAD/attention.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ce8755a66fa65637"}},{"code_sha256_prefix":"c93a3ed41b61c72f","entry":"dot_attention","repo":"bzhangGo/zero","repo_kind":"official","path":"modules/rela.py","file_url":"https://github.com/bzhangGo/zero/blob/HEAD/modules/rela.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"c93a3ed41b61c72f"}},{"code_sha256_prefix":"245861c41603a891","entry":"gated_rms_norm","repo":"bzhangGo/zero","repo_kind":"official","path":"modules/rela.py","file_url":"https://github.com/bzhangGo/zero/blob/HEAD/modules/rela.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"BSD-3-Clause","inline_ok":true,"mcp_get_code":{"code_sha256":"245861c41603a891"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}