{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XL6ORE4ZLAKWCFG57AUCUPJQJR","short_pith_number":"pith:XL6ORE4Z","schema_version":"1.0","canonical_sha256":"bafce8939958156114ddf8282a3d304c50f1a3dc91c6cd7f992620a24ed84eda","source":{"kind":"arxiv","id":"2402.15637","version":2},"attestation_state":"computed","paper":{"title":"Addressing Order Sensitivity of In-Context Demonstration Examples in Causal Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanqi Yan, Lin Gui, Yanzheng Xiang, Yulan He","submitted_at":"2024-02-23T22:39:12Z","abstract_excerpt":"In-context learning has become a popular paradigm in natural language processing. However, its performance can be significantly influenced by the order of in-context demonstration examples. In this paper, we found that causal language models (CausalLMs) are more sensitive to this order compared to prefix language models (PrefixLMs). We attribute this phenomenon to the auto-regressive attention masks within CausalLMs, which restrict each token from accessing information from subsequent tokens. This results in different receptive fields for samples at different positions, thereby leading to repr"},"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":"2402.15637","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-23T22:39:12Z","cross_cats_sorted":[],"title_canon_sha256":"95129d3289acad71be392f6db317be1795859b5032d4cba7606413a66caa8ee0","abstract_canon_sha256":"7897b79d30321bda65fcb12a847228040317dbcfdb447502d82e743910c46335"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:06.932427Z","signature_b64":"OMCvqbx5RttZGvHE3zULqo2+Tye8cHrafj+wN8gAGL2b/15ElAeqPEQEA17c45JYv7KcVD6o/r3MPh/8OicUCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bafce8939958156114ddf8282a3d304c50f1a3dc91c6cd7f992620a24ed84eda","last_reissued_at":"2026-07-05T08:28:06.931949Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:06.931949Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Addressing Order Sensitivity of In-Context Demonstration Examples in Causal Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanqi Yan, Lin Gui, Yanzheng Xiang, Yulan He","submitted_at":"2024-02-23T22:39:12Z","abstract_excerpt":"In-context learning has become a popular paradigm in natural language processing. However, its performance can be significantly influenced by the order of in-context demonstration examples. In this paper, we found that causal language models (CausalLMs) are more sensitive to this order compared to prefix language models (PrefixLMs). We attribute this phenomenon to the auto-regressive attention masks within CausalLMs, which restrict each token from accessing information from subsequent tokens. This results in different receptive fields for samples at different positions, thereby leading to repr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15637","kind":"arxiv","version":2},"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/2402.15637/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":"2402.15637","created_at":"2026-07-05T08:28:06.932005+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.15637v2","created_at":"2026-07-05T08:28:06.932005+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15637","created_at":"2026-07-05T08:28:06.932005+00:00"},{"alias_kind":"pith_short_12","alias_value":"XL6ORE4ZLAKW","created_at":"2026-07-05T08:28:06.932005+00:00"},{"alias_kind":"pith_short_16","alias_value":"XL6ORE4ZLAKWCFG5","created_at":"2026-07-05T08:28:06.932005+00:00"},{"alias_kind":"pith_short_8","alias_value":"XL6ORE4Z","created_at":"2026-07-05T08:28:06.932005+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.07936","citing_title":"Mimicking or Reasoning: Rethinking Multi-Modal In-Context Learning in Vision-Language Models","ref_index":34,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR","json":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR.json","graph_json":"https://pith.science/api/pith-number/XL6ORE4ZLAKWCFG57AUCUPJQJR/graph.json","events_json":"https://pith.science/api/pith-number/XL6ORE4ZLAKWCFG57AUCUPJQJR/events.json","paper":"https://pith.science/paper/XL6ORE4Z"},"agent_actions":{"view_html":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR","download_json":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR.json","view_paper":"https://pith.science/paper/XL6ORE4Z","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.15637&json=true","fetch_graph":"https://pith.science/api/pith-number/XL6ORE4ZLAKWCFG57AUCUPJQJR/graph.json","fetch_events":"https://pith.science/api/pith-number/XL6ORE4ZLAKWCFG57AUCUPJQJR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR/action/storage_attestation","attest_author":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR/action/author_attestation","sign_citation":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR/action/citation_signature","submit_replication":"https://pith.science/pith/XL6ORE4ZLAKWCFG57AUCUPJQJR/action/replication_record"}},"created_at":"2026-07-05T08:28:06.932005+00:00","updated_at":"2026-07-05T08:28:06.932005+00:00"}