{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RXWSNNWP26TOIBXQYVFYSGRWHZ","short_pith_number":"pith:RXWSNNWP","schema_version":"1.0","canonical_sha256":"8ded26b6cfd7a6e406f0c54b891a363e7068a8b9bd9343ce3823f520ef3d0d39","source":{"kind":"arxiv","id":"2205.12701","version":2},"attestation_state":"computed","paper":{"title":"Eliciting and Understanding Cross-Task Skills with Task-Level Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Juan Zha, Qinyuan Ye, Xiang Ren","submitted_at":"2022-05-25T11:59:05Z","abstract_excerpt":"Recent works suggest that transformer models are capable of multi-tasking on diverse NLP tasks and adapting to new tasks efficiently. However, the potential of these multi-task models may be limited as they use the same set of parameters for all tasks. In contrast, humans tackle tasks in a more flexible way, by making proper presumptions on what skills and knowledge are relevant and executing only the necessary computations. Inspired by this, we propose to use task-level mixture-of-expert models, which has a collection of transformer layers (i.e., experts) and a router component that chooses f"},"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":"2205.12701","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-05-25T11:59:05Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"0bbd94e363a4f6ae8625524640673f1c3518bd6008cb99c01db25ac6bef55016","abstract_canon_sha256":"411c397090621e114e3fd0fc6b2b0c3776ffbc40d184cf74a2eb8ed3b138420b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:18:03.849789Z","signature_b64":"zTY9qCAq7BpLzYIopFaSc3PjlfY5uROncW6x4EWgAVmtEUIHY/fil9LWrDfLqpaPsSPNWLVurzLrloz/v1bFCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8ded26b6cfd7a6e406f0c54b891a363e7068a8b9bd9343ce3823f520ef3d0d39","last_reissued_at":"2026-07-05T05:18:03.849376Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:18:03.849376Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Eliciting and Understanding Cross-Task Skills with Task-Level Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Juan Zha, Qinyuan Ye, Xiang Ren","submitted_at":"2022-05-25T11:59:05Z","abstract_excerpt":"Recent works suggest that transformer models are capable of multi-tasking on diverse NLP tasks and adapting to new tasks efficiently. However, the potential of these multi-task models may be limited as they use the same set of parameters for all tasks. In contrast, humans tackle tasks in a more flexible way, by making proper presumptions on what skills and knowledge are relevant and executing only the necessary computations. Inspired by this, we propose to use task-level mixture-of-expert models, which has a collection of transformer layers (i.e., experts) and a router component that chooses f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.12701","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/2205.12701/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":"2205.12701","created_at":"2026-07-05T05:18:03.849432+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.12701v2","created_at":"2026-07-05T05:18:03.849432+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.12701","created_at":"2026-07-05T05:18:03.849432+00:00"},{"alias_kind":"pith_short_12","alias_value":"RXWSNNWP26TO","created_at":"2026-07-05T05:18:03.849432+00:00"},{"alias_kind":"pith_short_16","alias_value":"RXWSNNWP26TOIBXQ","created_at":"2026-07-05T05:18:03.849432+00:00"},{"alias_kind":"pith_short_8","alias_value":"RXWSNNWP","created_at":"2026-07-05T05:18:03.849432+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.13468","citing_title":"An Interdisciplinary Approach to Human-Centered Machine Translation","ref_index":198,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ","json":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ.json","graph_json":"https://pith.science/api/pith-number/RXWSNNWP26TOIBXQYVFYSGRWHZ/graph.json","events_json":"https://pith.science/api/pith-number/RXWSNNWP26TOIBXQYVFYSGRWHZ/events.json","paper":"https://pith.science/paper/RXWSNNWP"},"agent_actions":{"view_html":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ","download_json":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ.json","view_paper":"https://pith.science/paper/RXWSNNWP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.12701&json=true","fetch_graph":"https://pith.science/api/pith-number/RXWSNNWP26TOIBXQYVFYSGRWHZ/graph.json","fetch_events":"https://pith.science/api/pith-number/RXWSNNWP26TOIBXQYVFYSGRWHZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ/action/storage_attestation","attest_author":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ/action/author_attestation","sign_citation":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ/action/citation_signature","submit_replication":"https://pith.science/pith/RXWSNNWP26TOIBXQYVFYSGRWHZ/action/replication_record"}},"created_at":"2026-07-05T05:18:03.849432+00:00","updated_at":"2026-07-05T05:18:03.849432+00:00"}