{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:DR3QNHMJLWWGL5WKPJP2FCHQGI","short_pith_number":"pith:DR3QNHMJ","canonical_record":{"source":{"id":"2406.20052","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T17:03:51Z","cross_cats_sorted":[],"title_canon_sha256":"62bc991d9352f9c5ba226f75663340bb0ab422a913e3fee339d651b4a076270b","abstract_canon_sha256":"e0e63cb4035e71bbaa5dc171699e15e8d9863bc836601ef8a645a02313471c43"},"schema_version":"1.0"},"canonical_sha256":"1c77069d895dac65f6ca7a5fa288f03210e436a90b3b113e3bf188460c81db13","source":{"kind":"arxiv","id":"2406.20052","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.20052","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"arxiv_version","alias_value":"2406.20052v3","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.20052","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"pith_short_12","alias_value":"DR3QNHMJLWWG","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"pith_short_16","alias_value":"DR3QNHMJLWWGL5WK","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"pith_short_8","alias_value":"DR3QNHMJ","created_at":"2026-07-05T10:44:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:DR3QNHMJLWWGL5WKPJP2FCHQGI","target":"record","payload":{"canonical_record":{"source":{"id":"2406.20052","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T17:03:51Z","cross_cats_sorted":[],"title_canon_sha256":"62bc991d9352f9c5ba226f75663340bb0ab422a913e3fee339d651b4a076270b","abstract_canon_sha256":"e0e63cb4035e71bbaa5dc171699e15e8d9863bc836601ef8a645a02313471c43"},"schema_version":"1.0"},"canonical_sha256":"1c77069d895dac65f6ca7a5fa288f03210e436a90b3b113e3bf188460c81db13","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:44:14.989725Z","signature_b64":"s1sYGSm2tCd6ikNse5hC+XjxffCuEPyzR5QR8M3BsAI0KeqaVvo3uawbogRFjnEtU79I+043nMi5eOhXOsifCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c77069d895dac65f6ca7a5fa288f03210e436a90b3b113e3bf188460c81db13","last_reissued_at":"2026-07-05T10:44:14.989243Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:44:14.989243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.20052","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:44:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NvOWZQxKP/fLKsJohfux8aQISRf9Y6MeqTPXkTnzq+qOVYI2HahneSVibp0kmZx3aTIXjE+7WmlvQelQMCJgCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-16T03:24:40.541832Z"},"content_sha256":"db51f5391c6cca5c7f24e206185e8c4e2afca393b2c8ba9f6cbfc33539b818c7","schema_version":"1.0","event_id":"sha256:db51f5391c6cca5c7f24e206185e8c4e2afca393b2c8ba9f6cbfc33539b818c7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:DR3QNHMJLWWGL5WKPJP2FCHQGI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding and Mitigating Language Confusion in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexandre B\\'erard, Kelly Marchisio, Sebastian Ruder, Th\\'eo Dehaze, Wei-Yin Ko","submitted_at":"2024-06-28T17:03:51Z","abstract_excerpt":"We investigate a surprising limitation of LLMs: their inability to consistently generate text in a user's desired language. We create the Language Confusion Benchmark (LCB) to evaluate such failures, covering 15 typologically diverse languages with existing and newly-created English and multilingual prompts. We evaluate a range of LLMs on monolingual and cross-lingual generation reflecting practical use cases, finding that Llama Instruct and Mistral models exhibit high degrees of language confusion and even the strongest models fail to consistently respond in the correct language. We observe t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.20052","kind":"arxiv","version":3},"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/2406.20052/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:44:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6z3OCGM5BhxTKbeXj3oUH0wTUgaZ05JXDZ3Yj7CvmyzIA4f00Ey7pr14AT5m9QfJoOYQH5Fo7NmuCeqFyuAQBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-16T03:24:40.542339Z"},"content_sha256":"52c5f3261f2d1ed95922e4c84fbd3bbf4234193a8848a6bf289883e71a109a02","schema_version":"1.0","event_id":"sha256:52c5f3261f2d1ed95922e4c84fbd3bbf4234193a8848a6bf289883e71a109a02"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI/bundle.json","state_url":"https://pith.science/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-16T03:24:40Z","links":{"resolver":"https://pith.science/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI","bundle":"https://pith.science/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI/bundle.json","state":"https://pith.science/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DR3QNHMJLWWGL5WKPJP2FCHQGI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:DR3QNHMJLWWGL5WKPJP2FCHQGI","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"e0e63cb4035e71bbaa5dc171699e15e8d9863bc836601ef8a645a02313471c43","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T17:03:51Z","title_canon_sha256":"62bc991d9352f9c5ba226f75663340bb0ab422a913e3fee339d651b4a076270b"},"schema_version":"1.0","source":{"id":"2406.20052","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.20052","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"arxiv_version","alias_value":"2406.20052v3","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.20052","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"pith_short_12","alias_value":"DR3QNHMJLWWG","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"pith_short_16","alias_value":"DR3QNHMJLWWGL5WK","created_at":"2026-07-05T10:44:14Z"},{"alias_kind":"pith_short_8","alias_value":"DR3QNHMJ","created_at":"2026-07-05T10:44:14Z"}],"graph_snapshots":[{"event_id":"sha256:52c5f3261f2d1ed95922e4c84fbd3bbf4234193a8848a6bf289883e71a109a02","target":"graph","created_at":"2026-07-05T10:44:14Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2406.20052/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We investigate a surprising limitation of LLMs: their inability to consistently generate text in a user's desired language. We create the Language Confusion Benchmark (LCB) to evaluate such failures, covering 15 typologically diverse languages with existing and newly-created English and multilingual prompts. We evaluate a range of LLMs on monolingual and cross-lingual generation reflecting practical use cases, finding that Llama Instruct and Mistral models exhibit high degrees of language confusion and even the strongest models fail to consistently respond in the correct language. We observe t","authors_text":"Alexandre B\\'erard, Kelly Marchisio, Sebastian Ruder, Th\\'eo Dehaze, Wei-Yin Ko","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T17:03:51Z","title":"Understanding and Mitigating Language Confusion in LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.20052","kind":"arxiv","version":3},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:db51f5391c6cca5c7f24e206185e8c4e2afca393b2c8ba9f6cbfc33539b818c7","target":"record","created_at":"2026-07-05T10:44:14Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"e0e63cb4035e71bbaa5dc171699e15e8d9863bc836601ef8a645a02313471c43","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T17:03:51Z","title_canon_sha256":"62bc991d9352f9c5ba226f75663340bb0ab422a913e3fee339d651b4a076270b"},"schema_version":"1.0","source":{"id":"2406.20052","kind":"arxiv","version":3}},"canonical_sha256":"1c77069d895dac65f6ca7a5fa288f03210e436a90b3b113e3bf188460c81db13","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1c77069d895dac65f6ca7a5fa288f03210e436a90b3b113e3bf188460c81db13","first_computed_at":"2026-07-05T10:44:14.989243Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:44:14.989243Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"s1sYGSm2tCd6ikNse5hC+XjxffCuEPyzR5QR8M3BsAI0KeqaVvo3uawbogRFjnEtU79I+043nMi5eOhXOsifCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:44:14.989725Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.20052","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:db51f5391c6cca5c7f24e206185e8c4e2afca393b2c8ba9f6cbfc33539b818c7","sha256:52c5f3261f2d1ed95922e4c84fbd3bbf4234193a8848a6bf289883e71a109a02"],"state_sha256":"5e32f8b1d83350734c06c577c61d3b9fe0512b9353461f19d45db7159603f3b3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0YaQcGXe/qYg7exxVqGl6aKewOQ7POlNWzM06oGiuR3B9EDCemqN/Y3SdoBuMMj6ppfkZq2BznXsbdrmo/ClBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-16T03:24:40.544828Z","bundle_sha256":"d3a77826ce5634dcbf8c53eacf55e4d77cd8127370ec499723c691659c579da6"}}