{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JGFNONAO4JFI7Q2FEPZ4MKMNEO","short_pith_number":"pith:JGFNONAO","schema_version":"1.0","canonical_sha256":"498ad7340ee24a8fc34523f3c6298d2386fc369eb3e8a09df1edf0669e3b333b","source":{"kind":"arxiv","id":"2504.02349","version":1},"attestation_state":"computed","paper":{"title":"Large (Vision) Language Models are Unsupervised In-Context Learners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Amir Zamir, Andrei Atanov, Artyom Gadetsky, Ghazal Hosseini Mighan, Maria Brbic, Yulun Jiang, Zhitong Gao","submitted_at":"2025-04-03T07:33:02Z","abstract_excerpt":"Recent advances in large language and vision-language models have enabled zero-shot inference, allowing models to solve new tasks without task-specific training. Various adaptation techniques such as prompt engineering, In-Context Learning (ICL), and supervised fine-tuning can further enhance the model's performance on a downstream task, but they require substantial manual effort to construct effective prompts or labeled examples. In this work, we introduce a joint inference framework for fully unsupervised adaptation, eliminating the need for manual prompt engineering and labeled examples. Un"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2504.02349","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-04-03T07:33:02Z","cross_cats_sorted":[],"title_canon_sha256":"2de426ee68fd78a269b0b9b0040fc69c8564d42cd397e46de60721e8f4076249","abstract_canon_sha256":"69eb0c75aecebaa9464b02bbba1d23672ca740e5e0bd2846d72bbab52a5060e1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:43:55.285171Z","signature_b64":"rORjzG0xyrtU+D8vrFTs+iCHHiNusB7v5h4sGMmZ9GFoJV/MZV9SXzGDgNA/jkQ3GTopYUCnXjXnSir4M5tUBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"498ad7340ee24a8fc34523f3c6298d2386fc369eb3e8a09df1edf0669e3b333b","last_reissued_at":"2026-07-05T10:43:55.284686Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:43:55.284686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large (Vision) Language Models are Unsupervised In-Context Learners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Amir Zamir, Andrei Atanov, Artyom Gadetsky, Ghazal Hosseini Mighan, Maria Brbic, Yulun Jiang, Zhitong Gao","submitted_at":"2025-04-03T07:33:02Z","abstract_excerpt":"Recent advances in large language and vision-language models have enabled zero-shot inference, allowing models to solve new tasks without task-specific training. Various adaptation techniques such as prompt engineering, In-Context Learning (ICL), and supervised fine-tuning can further enhance the model's performance on a downstream task, but they require substantial manual effort to construct effective prompts or labeled examples. In this work, we introduce a joint inference framework for fully unsupervised adaptation, eliminating the need for manual prompt engineering and labeled examples. Un"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.02349","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.02349/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2504.02349","created_at":"2026-07-05T10:43:55.284744+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.02349v1","created_at":"2026-07-05T10:43:55.284744+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.02349","created_at":"2026-07-05T10:43:55.284744+00:00"},{"alias_kind":"pith_short_12","alias_value":"JGFNONAO4JFI","created_at":"2026-07-05T10:43:55.284744+00:00"},{"alias_kind":"pith_short_16","alias_value":"JGFNONAO4JFI7Q2F","created_at":"2026-07-05T10:43:55.284744+00:00"},{"alias_kind":"pith_short_8","alias_value":"JGFNONAO","created_at":"2026-07-05T10:43:55.284744+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18089","citing_title":"From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20234","citing_title":"TabPFN-MT: A Natively Multitask In-Context Learner for Tabular Data","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO","json":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO.json","graph_json":"https://pith.science/api/pith-number/JGFNONAO4JFI7Q2FEPZ4MKMNEO/graph.json","events_json":"https://pith.science/api/pith-number/JGFNONAO4JFI7Q2FEPZ4MKMNEO/events.json","paper":"https://pith.science/paper/JGFNONAO"},"agent_actions":{"view_html":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO","download_json":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO.json","view_paper":"https://pith.science/paper/JGFNONAO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.02349&json=true","fetch_graph":"https://pith.science/api/pith-number/JGFNONAO4JFI7Q2FEPZ4MKMNEO/graph.json","fetch_events":"https://pith.science/api/pith-number/JGFNONAO4JFI7Q2FEPZ4MKMNEO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO/action/storage_attestation","attest_author":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO/action/author_attestation","sign_citation":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO/action/citation_signature","submit_replication":"https://pith.science/pith/JGFNONAO4JFI7Q2FEPZ4MKMNEO/action/replication_record"}},"created_at":"2026-07-05T10:43:55.284744+00:00","updated_at":"2026-07-05T10:43:55.284744+00:00"}