{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:PVRSATCQ4GVGWLAU2Z5NJC6PGD","short_pith_number":"pith:PVRSATCQ","schema_version":"1.0","canonical_sha256":"7d63204c50e1aa6b2c14d67ad48bcf30dc28a3cbc4551354453f5eb64e843856","source":{"kind":"arxiv","id":"2601.05366","version":2},"attestation_state":"computed","paper":{"title":"Lost in Execution: On the Multilingual Robustness of Tool Calling in Large Language Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Hua Wei, Ogochukwu N Okoani, T Pranav Kutralingam, Wanpeng Xu, Xiyang Hu, Zheng Luo","submitted_at":"2026-01-08T20:44:28Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly deployed as agents that invoke external tools through structured function calls. While recent work reports strong tool-calling performance under standard English-centric evaluations, the robustness of tool calling under multilingual user interactions remains underexplored. In this work, we introduce MLCL, a diagnostic benchmark, and conduct a systematic evaluation of multilingual tool calling across Chinese, Hindi, and the low-resource language Igbo. Through fine-grained error analysis, we show that many failures occur despite correct intent unders"},"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":"2601.05366","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-08T20:44:28Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"18a94cfbe5011e7e8c9adf5e4033e1d599b6eb03fc2dc23e52b40093adc3f4ca","abstract_canon_sha256":"e82c98a7ae8d79e7cc2962950eba777bad7f0feb8c05e7d5f1733630d77bf2f9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:17:32.378900Z","signature_b64":"WsI2JxCMACvFPQNgQo72rcLJ0rrUWjolB3pSfML7U9YWflgZiqEB7liu17jifdcr4z6iZQuMh7bz3RqFt2PWBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7d63204c50e1aa6b2c14d67ad48bcf30dc28a3cbc4551354453f5eb64e843856","last_reissued_at":"2026-06-30T01:17:32.377960Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:17:32.377960Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lost in Execution: On the Multilingual Robustness of Tool Calling in Large Language Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Hua Wei, Ogochukwu N Okoani, T Pranav Kutralingam, Wanpeng Xu, Xiyang Hu, Zheng Luo","submitted_at":"2026-01-08T20:44:28Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly deployed as agents that invoke external tools through structured function calls. While recent work reports strong tool-calling performance under standard English-centric evaluations, the robustness of tool calling under multilingual user interactions remains underexplored. In this work, we introduce MLCL, a diagnostic benchmark, and conduct a systematic evaluation of multilingual tool calling across Chinese, Hindi, and the low-resource language Igbo. Through fine-grained error analysis, we show that many failures occur despite correct intent unders"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.05366","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/2601.05366/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":"2601.05366","created_at":"2026-06-30T01:17:32.378088+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.05366v2","created_at":"2026-06-30T01:17:32.378088+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.05366","created_at":"2026-06-30T01:17:32.378088+00:00"},{"alias_kind":"pith_short_12","alias_value":"PVRSATCQ4GVG","created_at":"2026-06-30T01:17:32.378088+00:00"},{"alias_kind":"pith_short_16","alias_value":"PVRSATCQ4GVGWLAU","created_at":"2026-06-30T01:17:32.378088+00:00"},{"alias_kind":"pith_short_8","alias_value":"PVRSATCQ","created_at":"2026-06-30T01:17:32.378088+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2606.07017","citing_title":"The Sim-to-Real Gap of Foundation Model Agents: A Unified MDP Perspective","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2606.01936","citing_title":"What to Format and How: A Benchmark and Workflow Approach for Document Formatting","ref_index":37,"is_internal_anchor":true},{"citing_arxiv_id":"2604.24929","citing_title":"GAIA-v2-LILT: Multilingual Adaptation of Agent Benchmark beyond Translation","ref_index":5,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD","json":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD.json","graph_json":"https://pith.science/api/pith-number/PVRSATCQ4GVGWLAU2Z5NJC6PGD/graph.json","events_json":"https://pith.science/api/pith-number/PVRSATCQ4GVGWLAU2Z5NJC6PGD/events.json","paper":"https://pith.science/paper/PVRSATCQ"},"agent_actions":{"view_html":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD","download_json":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD.json","view_paper":"https://pith.science/paper/PVRSATCQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.05366&json=true","fetch_graph":"https://pith.science/api/pith-number/PVRSATCQ4GVGWLAU2Z5NJC6PGD/graph.json","fetch_events":"https://pith.science/api/pith-number/PVRSATCQ4GVGWLAU2Z5NJC6PGD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD/action/storage_attestation","attest_author":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD/action/author_attestation","sign_citation":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD/action/citation_signature","submit_replication":"https://pith.science/pith/PVRSATCQ4GVGWLAU2Z5NJC6PGD/action/replication_record"}},"created_at":"2026-06-30T01:17:32.378088+00:00","updated_at":"2026-06-30T01:17:32.378088+00:00"}