{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:Q36RX6HGW2X6AK3KOTLQMFSSTZ","short_pith_number":"pith:Q36RX6HG","schema_version":"1.0","canonical_sha256":"86fd1bf8e6b6afe02b6a74d70616529e73e341048a88de09a7bc13f4a80dc198","source":{"kind":"arxiv","id":"2109.01134","version":6},"attestation_state":"computed","paper":{"title":"Learning to Prompt for Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Chen Change Loy, Jingkang Yang, Kaiyang Zhou, Ziwei Liu","submitted_at":"2021-09-02T17:57:31Z","abstract_excerpt":"Large pre-trained vision-language models like CLIP have shown great potential in learning representations that are transferable across a wide range of downstream tasks. Different from the traditional representation learning that is based mostly on discretized labels, vision-language pre-training aligns images and texts in a common feature space, which allows zero-shot transfer to a downstream task via prompting, i.e., classification weights are synthesized from natural language describing classes of interest. In this work, we show that a major challenge for deploying such models in practice is"},"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":"2109.01134","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-09-02T17:57:31Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"6c42ea7e7a4948ba7de0e2274544cad22758471f804068c36283f7839520d136","abstract_canon_sha256":"71ba7909d4a0ed3a0527535fab7450fef50b8c2d495b5f0bdf111962cce42d57"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:03:49.610540Z","signature_b64":"BZ3SgnMt84Q5K/Sj/FUvLI6zo2FJBhJvArAoptWSesvglp4us1dFk6BH8aEPRpU0Unk94bWFMZrhlI7mLf7eDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"86fd1bf8e6b6afe02b6a74d70616529e73e341048a88de09a7bc13f4a80dc198","last_reissued_at":"2026-07-05T05:03:49.609923Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:03:49.609923Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning to Prompt for Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Chen Change Loy, Jingkang Yang, Kaiyang Zhou, Ziwei Liu","submitted_at":"2021-09-02T17:57:31Z","abstract_excerpt":"Large pre-trained vision-language models like CLIP have shown great potential in learning representations that are transferable across a wide range of downstream tasks. Different from the traditional representation learning that is based mostly on discretized labels, vision-language pre-training aligns images and texts in a common feature space, which allows zero-shot transfer to a downstream task via prompting, i.e., classification weights are synthesized from natural language describing classes of interest. In this work, we show that a major challenge for deploying such models in practice is"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.01134","kind":"arxiv","version":6},"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/2109.01134/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":"2109.01134","created_at":"2026-07-05T05:03:49.610005+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.01134v6","created_at":"2026-07-05T05:03:49.610005+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.01134","created_at":"2026-07-05T05:03:49.610005+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q36RX6HGW2X6","created_at":"2026-07-05T05:03:49.610005+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q36RX6HGW2X6AK3K","created_at":"2026-07-05T05:03:49.610005+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q36RX6HG","created_at":"2026-07-05T05:03:49.610005+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":16,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25657","citing_title":"Steering Vision-Language Models with Joint Sparse Autoencoders","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2208.14649","citing_title":"DetailCLIP: Injecting Image Details into CLIP's Feature Space","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17187","citing_title":"PluRule: A Benchmark for Moderating Pluralistic Communities on Social Media","ref_index":178,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18464","citing_title":"PERL: Parameter Efficient Reasoning in CLIP Latent Space","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18162","citing_title":"Self-Evolving Spatial Reasoning in Vision Language Models via Geometric Logic Consistency","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2512.20249","citing_title":"Unified Multimodal Brain Decoding via Cross-Subject Soft-ROI Fusion","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11572","citing_title":"TB-AVA: Text as a Semantic Bridge for Audio-Visual Parameter Efficient Finetuning","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12952","citing_title":"Debunking Grad-ECLIP: A Comprehensive Study on Its Incorrectness and Fundamental Principles for Model Interpretation","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11572","citing_title":"TB-AVA: Text as a Semantic Bridge for Audio-Visual Parameter Efficient Finetuning","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2208.01618","citing_title":"An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02912","citing_title":"Reasoning-Guided Grounding: Elevating Video Anomaly Detection through Multimodal Large Language Models","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18444","citing_title":"ProtoCLIP: Prototype-Aligned Latent Refinement for Robust Zero-Shot Chest X-Ray Classification","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20311","citing_title":"Seeing Further and Wider: Joint Spatio-Temporal Enlargement for Micro-Video Popularity Prediction","ref_index":81,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03294","citing_title":"FACTOR: Counterfactual Training-Free Test-Time Adaptation for Open-Vocabulary Object Detection","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06477","citing_title":"GeoStack: A Framework for Quasi-Abelian Knowledge Composition in VLMs","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02604","citing_title":"Rethinking the Need for Source Models: Source-Free Domain Adaptation from Scratch Guided by a Vision-Language Model","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ","json":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ.json","graph_json":"https://pith.science/api/pith-number/Q36RX6HGW2X6AK3KOTLQMFSSTZ/graph.json","events_json":"https://pith.science/api/pith-number/Q36RX6HGW2X6AK3KOTLQMFSSTZ/events.json","paper":"https://pith.science/paper/Q36RX6HG"},"agent_actions":{"view_html":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ","download_json":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ.json","view_paper":"https://pith.science/paper/Q36RX6HG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.01134&json=true","fetch_graph":"https://pith.science/api/pith-number/Q36RX6HGW2X6AK3KOTLQMFSSTZ/graph.json","fetch_events":"https://pith.science/api/pith-number/Q36RX6HGW2X6AK3KOTLQMFSSTZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ/action/storage_attestation","attest_author":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ/action/author_attestation","sign_citation":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ/action/citation_signature","submit_replication":"https://pith.science/pith/Q36RX6HGW2X6AK3KOTLQMFSSTZ/action/replication_record"}},"created_at":"2026-07-05T05:03:49.610005+00:00","updated_at":"2026-07-05T05:03:49.610005+00:00"}