{"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/berts-are-generative-in-context-learners","title":"BERTs are Generative In-Context Learners","arxiv_id":"2406.04823","date":"2024-06-07","proceeding":null,"authors":["David Samuel"],"abstract":"While in-context learning is commonly associated with causal language models, such as GPT, we demonstrate that this capability also 'emerges' in masked language models. Through an embarrassingly simple inference technique, we enable an existing masked model, DeBERTa, to perform generative tasks without additional training or architectural changes. Our evaluation reveals that the masked and causal language models behave very differently, as they clearly outperform each other on different categories of tasks. These complementary strengths suggest that the field's focus on causal models for in-context learning may be limiting - both architectures can develop these capabilities, but with distinct advantages; pointing toward promising hybrid approaches that combine the strengths of both objectives.","url_abs":"https://arxiv.org/abs/2406.04823v2","url_pdf":"https://arxiv.org/pdf/2406.04823v2.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":"berts-are-generative-in-context-learners","repo_url":"https://github.com/ltgoslo/bert-in-context","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"in-context-learning","task_name":"In-Context Learning"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"cosine-annealing","method_name":"Cosine Annealing"},{"method_slug":"deberta","method_name":"DeBERTa"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"discriminative-fine-tuning","method_name":"Discriminative Fine-Tuning"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"focus","method_name":"Focus"},{"method_slug":"gpt","method_name":"GPT"},{"method_slug":"gpt-3","method_name":"GPT-3"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-cosine-annealing","method_name":"Linear Warmup With Cosine Annealing"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2406.04823","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2406.04823"}},"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/ltgoslo/bert-in-context","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"ran_violates":1,"ran":11,"ran_draft_wrong":1,"unverified":13},"by_repo_kind":{"official":{"samples":26,"ran":13,"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":"f6c275d6a18330a9","entry":"exact_match_score","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/record.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/record.py","link_basis":"plan_row","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f6c275d6a18330a9"}},{"code_sha256_prefix":"d9296df040618579","entry":"f1_score","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/record.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/record.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"d9296df040618579"}},{"code_sha256_prefix":"28110ec97c757855","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/hellaswag.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/hellaswag.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"28110ec97c757855"}},{"code_sha256_prefix":"11065df7f8515818","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/story_cloze.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/story_cloze.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"11065df7f8515818"}},{"code_sha256_prefix":"522de10ce348cb0c","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/winograd.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/winograd.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"522de10ce348cb0c"}},{"code_sha256_prefix":"8131636afc6ecc7d","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/glue.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/glue.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"8131636afc6ecc7d"}},{"code_sha256_prefix":"1edc73f4d0b284d9","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/multirc.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/multirc.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"1edc73f4d0b284d9"}},{"code_sha256_prefix":"67264be51677507e","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"needle-in-a-haystack/haystack.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/needle-in-a-haystack/haystack.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"67264be51677507e"}},{"code_sha256_prefix":"f423eb95fad23611","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/arc.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/arc.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f423eb95fad23611"}},{"code_sha256_prefix":"4f852be96a94479f","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/natural_questions.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/natural_questions.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"4f852be96a94479f"}},{"code_sha256_prefix":"41aac6a681c0af69","entry":"format_prompt","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/openbookqa.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/openbookqa.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"41aac6a681c0af69"}},{"code_sha256_prefix":"c6a80c065d2e4851","entry":"normalize_answer","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/record.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/record.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"c6a80c065d2e4851"}},{"code_sha256_prefix":"a4e00754dd4bcfaa","entry":"sample_random_examples","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/natural_questions.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/natural_questions.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"a4e00754dd4bcfaa"}},{"code_sha256_prefix":"86c6f499c0a026ce","entry":"get_matrix","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"needle-in-a-haystack/haystack_plot.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/needle-in-a-haystack/haystack_plot.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"86c6f499c0a026ce"}},{"code_sha256_prefix":"3c0f2aa1a66632c5","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/hellaswag.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/hellaswag.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3c0f2aa1a66632c5"}},{"code_sha256_prefix":"3bd35418b3afc655","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/story_cloze.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/story_cloze.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3bd35418b3afc655"}},{"code_sha256_prefix":"ea61f50fcaa11638","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/winograd.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/winograd.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"ea61f50fcaa11638"}},{"code_sha256_prefix":"414bac79226f48f5","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/winogrande.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/winogrande.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"414bac79226f48f5"}},{"code_sha256_prefix":"dd674d997564c62d","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/glue.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/glue.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"dd674d997564c62d"}},{"code_sha256_prefix":"79d613cc4f9f531b","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"glue/multirc.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/glue/multirc.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"79d613cc4f9f531b"}},{"code_sha256_prefix":"1866651164e5bad9","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"needle-in-a-haystack/haystack.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/needle-in-a-haystack/haystack.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"1866651164e5bad9"}},{"code_sha256_prefix":"7fbad6dad736d00b","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/arc.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/arc.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"7fbad6dad736d00b"}},{"code_sha256_prefix":"97a40c18e9c30df4","entry":"load_data","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/openbookqa.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/openbookqa.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"97a40c18e9c30df4"}},{"code_sha256_prefix":"081ead69959a12d1","entry":"load_model","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"language-modeling/hellaswag.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/language-modeling/hellaswag.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"081ead69959a12d1"}},{"code_sha256_prefix":"3585e698dbdb0fc1","entry":"load_model","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"needle-in-a-haystack/haystack.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/needle-in-a-haystack/haystack.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"3585e698dbdb0fc1"}},{"code_sha256_prefix":"03857b5d2df81f28","entry":"load_model","repo":"ltgoslo/bert-in-context","repo_kind":"official","path":"question-answering/natural_questions.py","file_url":"https://github.com/ltgoslo/bert-in-context/blob/HEAD/question-answering/natural_questions.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"03857b5d2df81f28"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}