{"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/discovery-of-natural-language-concepts-in","title":"Discovery of Natural Language Concepts in Individual Units of CNNs","arxiv_id":"1902.07249","date":"2019-02-18","proceeding":"ICLR 2019 5","authors":["Seil Na","Yo Joong Choe","Dong-Hyun Lee","Gunhee Kim"],"abstract":"Although deep convolutional networks have achieved improved performance in\nmany natural language tasks, they have been treated as black boxes because they\nare difficult to interpret. Especially, little is known about how they\nrepresent language in their intermediate layers. In an attempt to understand\nthe representations of deep convolutional networks trained on language tasks,\nwe show that individual units are selectively responsive to specific morphemes,\nwords, and phrases, rather than responding to arbitrary and uninterpretable\npatterns. In order to quantitatively analyze such an intriguing phenomenon, we\npropose a concept alignment method based on how units respond to the replicated\ntext. We conduct analyses with different architectures on multiple datasets for\nclassification and translation tasks and provide new insights into how deep\nmodels understand natural language.","url_abs":"http://arxiv.org/abs/1902.07249v2","url_pdf":"http://arxiv.org/pdf/1902.07249v2.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":"discovery-of-natural-language-concepts-in","repo_url":"https://github.com/seilna/CNN-Units-in-NLP","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"concept-alignment","task_name":"Concept Alignment"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1902.07249","atlas_url":"https://app.syntology.ai/?focus=1902.07249","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1902.07249"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/seilna/CNN-Units-in-NLP","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":3},"by_repo_kind":{"official":{"samples":3,"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":"8198392778043324","entry":"html_per_tas","repo":"seilna/CNN-Units-in-NLP","repo_kind":"official","path":"code/utils.py","file_url":"https://github.com/seilna/CNN-Units-in-NLP/blob/HEAD/code/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":"8198392778043324"}},{"code_sha256_prefix":"31e27ca4e46dca13","entry":"html_per_unit","repo":"seilna/CNN-Units-in-NLP","repo_kind":"official","path":"code/utils.py","file_url":"https://github.com/seilna/CNN-Units-in-NLP/blob/HEAD/code/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":"31e27ca4e46dca13"}},{"code_sha256_prefix":"889fdbb32959107f","entry":"sample_top","repo":"seilna/CNN-Units-in-NLP","repo_kind":"official","path":"code/utils.py","file_url":"https://github.com/seilna/CNN-Units-in-NLP/blob/HEAD/code/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":"889fdbb32959107f"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}