{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BCIHT6YR5WW6DWDFTT3MXWGREQ","short_pith_number":"pith:BCIHT6YR","schema_version":"1.0","canonical_sha256":"089079fb11edade1d8659cf6cbd8d1243d22f95a64f402f306c0f3e47c8d5b04","source":{"kind":"arxiv","id":"2605.27649","version":1},"attestation_state":"computed","paper":{"title":"Disentangling Language Roles in Multilingual LLM Task Execution","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Guansu Wang, Jiaxin Liu, Lei Zhao, Liang He, Man Liang, Minxuan Hu, Qishi Zhan, Seoyeon Jang, Xinyue Xiang, Ziheng Chen","submitted_at":"2026-05-26T20:09:34Z","abstract_excerpt":"Multilingual LLMs are increasingly used when instruction, source content, and required response languages do not coincide. Existing benchmarks have expanded multilingual instruction-following evaluation, but they rarely isolate these three roles within a fully crossed design. We introduce MTM-Bench, a controlled benchmark for language-conditioned task execution in which each instance is defined by a triplet \\((L_{\\text{instr}}, L_{\\text{content}}, L_{\\text{resp}})\\). Across English, Spanish, and Chinese, MTM-Bench enumerates all 27 triplets and contains 2{,}430 instances per model across seman"},"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":"2605.27649","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-26T20:09:34Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"c7cd7b916a28e460684ce93124d9d15937e1ccb0dc229784b165775e896d9d42","abstract_canon_sha256":"7df6cb096e0c498006289c4deee5d340e47ea91dc152c97136622330b72785eb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:45.335291Z","signature_b64":"KgvNkfxGT/2zlBwBEJ3nnaOtHpfUVNrGTmQRqjYQxaDfSgyiclMlq7Fgdvl+BQJLnsm01IDpjEuPy0O/qSjmBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"089079fb11edade1d8659cf6cbd8d1243d22f95a64f402f306c0f3e47c8d5b04","last_reissued_at":"2026-05-28T01:04:45.334823Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:45.334823Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Disentangling Language Roles in Multilingual LLM Task Execution","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Guansu Wang, Jiaxin Liu, Lei Zhao, Liang He, Man Liang, Minxuan Hu, Qishi Zhan, Seoyeon Jang, Xinyue Xiang, Ziheng Chen","submitted_at":"2026-05-26T20:09:34Z","abstract_excerpt":"Multilingual LLMs are increasingly used when instruction, source content, and required response languages do not coincide. Existing benchmarks have expanded multilingual instruction-following evaluation, but they rarely isolate these three roles within a fully crossed design. We introduce MTM-Bench, a controlled benchmark for language-conditioned task execution in which each instance is defined by a triplet \\((L_{\\text{instr}}, L_{\\text{content}}, L_{\\text{resp}})\\). Across English, Spanish, and Chinese, MTM-Bench enumerates all 27 triplets and contains 2{,}430 instances per model across seman"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.27649","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/2605.27649/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":"2605.27649","created_at":"2026-05-28T01:04:45.334878+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.27649v1","created_at":"2026-05-28T01:04:45.334878+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.27649","created_at":"2026-05-28T01:04:45.334878+00:00"},{"alias_kind":"pith_short_12","alias_value":"BCIHT6YR5WW6","created_at":"2026-05-28T01:04:45.334878+00:00"},{"alias_kind":"pith_short_16","alias_value":"BCIHT6YR5WW6DWDF","created_at":"2026-05-28T01:04:45.334878+00:00"},{"alias_kind":"pith_short_8","alias_value":"BCIHT6YR","created_at":"2026-05-28T01:04:45.334878+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ","json":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ.json","graph_json":"https://pith.science/api/pith-number/BCIHT6YR5WW6DWDFTT3MXWGREQ/graph.json","events_json":"https://pith.science/api/pith-number/BCIHT6YR5WW6DWDFTT3MXWGREQ/events.json","paper":"https://pith.science/paper/BCIHT6YR"},"agent_actions":{"view_html":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ","download_json":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ.json","view_paper":"https://pith.science/paper/BCIHT6YR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.27649&json=true","fetch_graph":"https://pith.science/api/pith-number/BCIHT6YR5WW6DWDFTT3MXWGREQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BCIHT6YR5WW6DWDFTT3MXWGREQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ/action/storage_attestation","attest_author":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ/action/author_attestation","sign_citation":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ/action/citation_signature","submit_replication":"https://pith.science/pith/BCIHT6YR5WW6DWDFTT3MXWGREQ/action/replication_record"}},"created_at":"2026-05-28T01:04:45.334878+00:00","updated_at":"2026-05-28T01:04:45.334878+00:00"}