{"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/large-concept-models-language-modeling-in-a","title":"Large Concept Models: Language Modeling in a Sentence Representation Space","arxiv_id":"2412.08821","date":"2024-12-11","proceeding":null,"authors":["LCM team","Loïc Barrault","Paul-Ambroise Duquenne","Maha Elbayad","Artyom Kozhevnikov","Belen Alastruey","Pierre Andrews","Mariano Coria","Guillaume Couairon","Marta R. Costa-jussà","David Dale","Hady Elsahar","Kevin Heffernan","João Maria Janeiro","Tuan Tran","Christophe Ropers","Eduardo Sánchez","Robin San Roman","Alexandre Mourachko","Safiyyah Saleem","Holger Schwenk"],"abstract":"LLMs have revolutionized the field of artificial intelligence and have emerged as the de-facto tool for many tasks. The current established technology of LLMs is to process input and generate output at the token level. This is in sharp contrast to humans who operate at multiple levels of abstraction, well beyond single words, to analyze information and to generate creative content. In this paper, we present an attempt at an architecture which operates on an explicit higher-level semantic representation, which we name a concept. Concepts are language- and modality-agnostic and represent a higher level idea or action in a flow. Hence, we build a \"Large Concept Model\". In this study, as proof of feasibility, we assume that a concept corresponds to a sentence, and use an existing sentence embedding space, SONAR, which supports up to 200 languages in both text and speech modalities. The Large Concept Model is trained to perform autoregressive sentence prediction in an embedding space. We explore multiple approaches, namely MSE regression, variants of diffusion-based generation, and models operating in a quantized SONAR space. These explorations are performed using 1.6B parameter models and training data in the order of 1.3T tokens. We then scale one architecture to a model size of 7B parameters and training data of about 2.7T tokens. We perform an experimental evaluation on several generative tasks, namely summarization and a new task of summary expansion. Finally, we show that our model exhibits impressive zero-shot generalization performance to many languages, outperforming existing LLMs of the same size. The training code of our models is freely available.","url_abs":"https://arxiv.org/abs/2412.08821v2","url_pdf":"https://arxiv.org/pdf/2412.08821v2.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":"large-concept-models-language-modeling-in-a","repo_url":"https://github.com/facebookresearch/large_concept_model","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"sentence-embedding","task_name":"Sentence Embedding"},{"task_slug":"sentence-embedding-1","task_name":"Sentence-Embedding"},{"task_slug":"zero-shot-generalization","task_name":"Zero-shot Generalization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2412.08821","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2412.08821"}},"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/facebookresearch/large_concept_model","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":10},"by_repo_kind":{"official":{"samples":10,"ran":0,"repositories":1}},"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":"dcc4c0f743e740f8","entry":"define_parquet_dataset","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/parquet_utils.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/parquet_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"dcc4c0f743e740f8"}},{"code_sha256_prefix":"69d2f7ac6f37cc44","entry":"filter_empty_string","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/sentence_splitting.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/sentence_splitting.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"69d2f7ac6f37cc44"}},{"code_sha256_prefix":"13a2a85ebdf50b7d","entry":"find_closest_indices","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/sentence_splitter_pipeline.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/sentence_splitter_pipeline.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"13a2a85ebdf50b7d"}},{"code_sha256_prefix":"3fea6348a0d357ee","entry":"get_renaming_mappers","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/configs.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/configs.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3fea6348a0d357ee"}},{"code_sha256_prefix":"a77dd5b86cb93ecb","entry":"insert_elements","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/sentence_splitter_pipeline.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/sentence_splitter_pipeline.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a77dd5b86cb93ecb"}},{"code_sha256_prefix":"a99d1b23e0f7c52e","entry":"merge_small_intervals","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/sentence_splitter_pipeline.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/sentence_splitter_pipeline.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a99d1b23e0f7c52e"}},{"code_sha256_prefix":"55e3f0eada7f40a3","entry":"remove_emojis","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/sentence_splitting.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/sentence_splitting.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"55e3f0eada7f40a3"}},{"code_sha256_prefix":"15193a94d33519a0","entry":"remove_non_printable_chars","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/sentence_splitting.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/sentence_splitting.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"15193a94d33519a0"}},{"code_sha256_prefix":"8c1c048b8f2f8356","entry":"return_none_on_failure","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/dataloading.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/dataloading.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8c1c048b8f2f8356"}},{"code_sha256_prefix":"0b0a924396ab6c1a","entry":"truncate_sequence","repo":"facebookresearch/large_concept_model","repo_kind":"official","path":"lcm/datasets/dataloader.py","file_url":"https://github.com/facebookresearch/large_concept_model/blob/HEAD/lcm/datasets/dataloader.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0b0a924396ab6c1a"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}