{"about":{"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.","site":"https://codewithpapers.app","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","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/method/grouped-query-attention","slug":"grouped-query-attention","name":"Grouped-query attention","full_name":"Grouped-query attention","full_name_withheld":false,"description_markdown":"**Grouped-query attention** an interpolation of multi-query and multi-head attention that achieves quality close to multi-head at comparable speed to multi-query attention.","description_state":"present","introduced_year":null,"introduced_by":{"title":null,"paper":null,"first_author":null,"n_authors":0,"url_abs":null,"archive_paper_url":null},"source":{"url":"https://arxiv.org/abs/2305.13245v3","title":"GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints","url_on_a_paper_host":true},"code_snippet_url":null,"code_snippet_url_on_a_code_host":false,"categories":[{"area":"General","area_id":"general","collection":"Attention","url":"/methods/category/attention","pwc_aliases":[]}],"n_papers_tagged":19,"archive_num_papers":null,"papers_newest_first":[{"paper":"/paper/hardware-efficient-attention-for-fast","slug":"hardware-efficient-attention-for-fast","title":"Hardware-Efficient Attention for Fast Decoding","date":"2025-05-27","arxiv_id":"2505.21487","n_code_links":2,"syntology":null},{"paper":null,"slug":"hunyuan-turbos-advancing-large-language","title":"Hunyuan-TurboS: Advancing Large Language Models through Mamba-Transformer Synergy and Adaptive Chain-of-Thought","date":"2025-05-21","arxiv_id":"2505.15431","n_code_links":0,"syntology":null},{"paper":"/paper/towards-economical-inference-enabling","slug":"towards-economical-inference-enabling","title":"Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs","date":"2025-02-20","arxiv_id":"2502.14837","n_code_links":1,"syntology":{"ran":10,"of":15,"n_ran_checked":10,"n_instrument":0,"unverified":5,"pointer_only":0,"phrase":"10 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 0 where Syntology's instrument failed) · 5 unverified","official":{"repos":["JT-Ushio/MHA2MLA"],"state":"official (archive's flag): 10 ran","n_ran":10,"n_constructed":0,"n_ran_no_instrument_failure":10,"n_unverified":5,"ran_from_kinds":["official"]}}},{"paper":"/paper/fastkv-kv-cache-compression-for-fast-long","slug":"fastkv-kv-cache-compression-for-fast-long","title":"FastKV: KV Cache Compression for Fast Long-Context Processing with Token-Selective Propagation","date":"2025-02-03","arxiv_id":"2502.01068","n_code_links":1,"syntology":null},{"paper":null,"slug":"sigma-differential-rescaling-of-query-key-and","title":"Sigma: Differential Rescaling of Query, Key and Value for Efficient Language Models","date":"2025-01-23","arxiv_id":"2501.13629","n_code_links":0,"syntology":null},{"paper":null,"slug":"uncomp-uncertainty-aware-long-context","title":"UNComp: Uncertainty-Aware Long-Context Compressor for Efficient Large Language Model Inference","date":"2024-10-04","arxiv_id":"2410.03090","n_code_links":0,"syntology":null},{"paper":null,"slug":"weighted-grouped-query-attention-in","title":"Weighted Grouped Query Attention in Transformers","date":"2024-07-15","arxiv_id":"2407.10855","n_code_links":0,"syntology":null},{"paper":"/paper/mlkv-multi-layer-key-value-heads-for-memory","slug":"mlkv-multi-layer-key-value-heads-for-memory","title":"MLKV: Multi-Layer Key-Value Heads for Memory Efficient Transformer Decoding","date":"2024-06-13","arxiv_id":"2406.09297","n_code_links":1,"syntology":null},{"paper":null,"slug":"effectively-compress-kv-heads-for-llm","title":"Effectively Compress KV Heads for LLM","date":"2024-06-11","arxiv_id":"2406.07056","n_code_links":0,"syntology":null},{"paper":"/paper/samba-simple-hybrid-state-space-models-for","slug":"samba-simple-hybrid-state-space-models-for","title":"Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling","date":"2024-06-11","arxiv_id":"2406.07522","n_code_links":2,"syntology":{"ran":1,"of":1,"n_ran_checked":1,"n_instrument":0,"unverified":0,"pointer_only":0,"phrase":"1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified; the one sample that ran constructed an object rather than computing a result","official":{"repos":["microsoft/Samba","sustcsonglin/flash-linear-attention"],"state":"official (archive's flag): 1 ran","n_ran":1,"n_constructed":1,"n_ran_no_instrument_failure":1,"n_unverified":0,"ran_from_kinds":["official"]}}},{"paper":"/paper/reducing-transformer-key-value-cache-size","slug":"reducing-transformer-key-value-cache-size","title":"Reducing Transformer Key-Value Cache Size with Cross-Layer Attention","date":"2024-05-21","arxiv_id":"2405.12981","n_code_links":2,"syntology":null},{"paper":null,"slug":"codeshell-technical-report","title":"CodeShell Technical Report","date":"2024-03-23","arxiv_id":"2403.15747","n_code_links":0,"syntology":null},{"paper":null,"slug":"vi-mistral-x-building-a-vietnamese-language","title":"Vi-Mistral-X: Building a Vietnamese Language Model with Advanced Continual Pre-training","date":"2024-03-20","arxiv_id":"2403.15470","n_code_links":0,"syntology":null},{"paper":"/paper/dynamic-memory-compression-retrofitting-llms","slug":"dynamic-memory-compression-retrofitting-llms","title":"Dynamic Memory Compression: Retrofitting LLMs for Accelerated Inference","date":"2024-03-14","arxiv_id":"2403.09636","n_code_links":1,"syntology":null},{"paper":"/paper/mobilellm-optimizing-sub-billion-parameter","slug":"mobilellm-optimizing-sub-billion-parameter","title":"MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases","date":"2024-02-22","arxiv_id":"2402.14905","n_code_links":4,"syntology":{"ran":11,"of":28,"n_ran_checked":10,"n_instrument":1,"unverified":17,"pointer_only":20,"phrase":"11 ran (of which 6 constructed an object rather than computing a result; 10 with no instrument failure: 1 honoured, 0 violated, 9 with no contract checked; 1 where Syntology's instrument failed) · 17 unverified","official":{"repos":["facebookresearch/mobilellm"],"state":"official (archive's flag): 6 ran","n_ran":6,"n_constructed":3,"n_ran_no_instrument_failure":6,"n_unverified":4,"ran_from_kinds":["listed","official"]}}},{"paper":"/paper/mistral-7b","slug":"mistral-7b","title":"Mistral 7B","date":"2023-10-10","arxiv_id":"2310.06825","n_code_links":6,"syntology":{"ran":10,"of":11,"n_ran_checked":10,"n_instrument":0,"unverified":1,"pointer_only":2,"phrase":"10 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 0 violated, 10 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified","official":{"repos":["mistralai/mistral-src"],"state":"official (archive's flag): 9 ran","n_ran":9,"n_constructed":0,"n_ran_no_instrument_failure":9,"n_unverified":1,"ran_from_kinds":["listed","official"]}}},{"paper":"/paper/llama-2-open-foundation-and-fine-tuned-chat","slug":"llama-2-open-foundation-and-fine-tuned-chat","title":"Llama 2: Open Foundation and Fine-Tuned Chat Models","date":"2023-07-18","arxiv_id":"2307.09288","n_code_links":19,"syntology":{"ran":33,"of":52,"n_ran_checked":20,"n_instrument":13,"unverified":19,"pointer_only":20,"phrase":"33 ran (of which 7 constructed an object rather than computing a result; 20 with no instrument failure: 1 honoured, 1 violated, 18 with no contract checked; 13 where Syntology's instrument failed) · 19 unverified","official":{"repos":["facebookresearch/llama"],"state":"community repositories only","n_ran":0,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["listed"]}}},{"paper":"/paper/gqa-training-generalized-multi-query","slug":"gqa-training-generalized-multi-query","title":"GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints","date":"2023-05-22","arxiv_id":"2305.13245","n_code_links":4,"syntology":{"ran":4,"of":5,"n_ran_checked":3,"n_instrument":1,"unverified":1,"pointer_only":3,"phrase":"4 ran (of which 2 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified","official":null}},{"paper":"/paper/flashattention-fast-and-memory-efficient","slug":"flashattention-fast-and-memory-efficient","title":"FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness","date":"2022-05-27","arxiv_id":"2205.14135","n_code_links":13,"syntology":{"ran":24,"of":30,"n_ran_checked":18,"n_instrument":6,"unverified":6,"pointer_only":1,"phrase":"24 ran (of which 2 constructed an object rather than computing a result; 18 with no instrument failure: 0 honoured, 1 violated, 17 with no contract checked; 6 where Syntology's instrument failed) · 6 unverified","official":{"repos":["dao-ailab/flash-attention"],"state":"official: no sample here; runs from other or unrecorded repositories","n_ran":0,"n_constructed":0,"n_ran_no_instrument_failure":0,"n_unverified":0,"ran_from_kinds":["found_in_text","listed"]}}}],"papers_shown":19,"papers_order":"date (newest first), then slug","paper_counts":{"archive_papers_tagged":19,"with_a_code_link":12,"where_syntology_ran_a_sample":7,"not_listed_spam_title":0,"listed":19,"listed_where_code_ran":7,"where_syntology_ran_a_sample_split":{"with_a_run_with_no_instrument_failure":7,"every_run_a_failure_of_syntologys_instrument":0,"listed_with_a_run_with_no_instrument_failure":7,"listed_every_run_a_failure_of_syntologys_instrument":0,"filter":{"states":["a run with no instrument failure","any run, instrument failures included"],"default":"a run with no instrument failure","note":"on the 'only where code ran' pages the default hides, in the browser, the rows where every run was a failure of Syntology's instrument; the second state shows them again. Rows are hidden, never re-ordered; these twins list every row"}},"definition":"distinct papers the archive tags; 'where Syntology ran a sample' counts papers with at least one harvested sample that ran, which is not a correctness claim"},"papers_where_code_ran":7,"papers_where_code_ran_first_page":"/method/grouped-query-attention/papers/ran/1","papers_where_code_ran_caption":"We ran code from the paper's repository; we did not isolate this method inside it.","pages":[{"order":"ran","page":1,"of":1,"url":"/method/grouped-query-attention/papers/ran/1","json":"/method/grouped-query-attention/papers/ran/1.json","rows":7}],"tasks":[{"task":"/task/language-modelling","name":"Language Modelling","papers":6},{"task":"/task/language-modeling","name":"Language Modeling","papers":5},{"task":"/task/large-language-model","name":"Large Language Model","papers":3},{"task":"/task/question-answering","name":"Question Answering","papers":3},{"task":"/task/4k","name":"4k","papers":2},{"task":"/task/arithmetic-reasoning","name":"Arithmetic Reasoning","papers":2},{"task":"/task/chatbot","name":"Chatbot","papers":2},{"task":"/task/code-generation","name":"Code Generation","papers":2},{"task":"/task/decoder","name":"Decoder","papers":2},{"task":null,"name":"GPU","papers":2},{"task":"/task/mamba","name":"Mamba","papers":2},{"task":"/task/math-word-problem-solving","name":"Math Word Problem Solving","papers":2},{"task":"/task/multi-task-language-understanding","name":"Multi-task Language Understanding","papers":2},{"task":"/task/sentence-completion","name":"Sentence Completion","papers":2},{"task":"/task/16k","name":"16k","papers":1},{"task":null,"name":"8k","papers":1},{"task":"/task/common-sense-reasoning","name":"Common Sense Reasoning","papers":1},{"task":"/task/computational-efficiency","name":"Computational Efficiency","papers":1},{"task":"/task/document-classification","name":"Document Classification","papers":1},{"task":"/task/humaneval","name":"HumanEval","papers":1}],"tasks_shown":20,"n_tasks":35,"usage_by_year":[{"year":"2022","papers":1},{"year":"2023","papers":3},{"year":"2024","papers":10},{"year":"2025","papers":5}],"row_source":"embedded","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/grouped-query-attention"},"syntology_read_at":"2026-10-05T00:36:11+00:00","syntology":{"source":"Syntology's have check for this name (the free coverage lookup), read from Syntology's graph","have_url":"https://syntology.ai/api/graph/have?x=Grouped-query%20attention","read_at":"2026-10-05T00:57:20+00:00","query":"Grouped-query attention","kind":"method","have":"unknown_name","tier":null,"verified_implementation":null,"level":null,"level_meaning":"the served verification level, per answer; never summed across methods","resolved_to":null,"resolution":null,"resolved_name_differs":false,"absence_asserted":false,"absence_unchecked":null,"no_code_is_about_this_method":false,"subject":null,"archive_paper_arxiv_id":null,"repositories":null,"retrieve_with":null,"name_competition":{"verdict":"none","may_be_wrong_sense":false,"bearers_found":0,"served":null,"leading_other_bearer":null,"n_other_bearers_listed":0,"note":"Nothing in this corpus is named 'Grouped-query attention' -- no method node, no paper announcing it in its title, no expanded form declaring it as an acronym."},"name_competition_text":null,"near_matches":["Grouped-Query Attention (GQA)","Cost-Optimal Grouped-Query Attention","Fuzzy Query Attention","Grouped Head Attention","Grouped Query Experts (GQE)","Grouped Self-Attention (GSA)","Grouped Cross Attention (GCA)","fixed-query oscillator attention"],"code_nearby":[],"code_nearby_caption":null,"origin_paper":null,"line":"name not resolved (unknown_name). The lookup matched no method, paper or venue by this name; that asserts nothing about whether code exists.","level_legend":null},"syntology_have_state":"loaded","record_sha256":"85af4d1f42a8f9abd9768151249aa31a4da602fc1bbea37603f18e60c00d07ce","record_changed_at":"2026-10-05","record_changed_at_basis":"changed"}